U.S. patent application number 15/399972 was filed with the patent office on 2017-07-13 for systems methods circuits and associated computer executable code for digital catalog augmentation.
The applicant listed for this patent is Klevu Oy. Invention is credited to Niraj Aswani, Nilay Oza.
Application Number | 20170200207 15/399972 |
Document ID | / |
Family ID | 59275750 |
Filed Date | 2017-07-13 |
United States Patent
Application |
20170200207 |
Kind Code |
A1 |
Aswani; Niraj ; et
al. |
July 13, 2017 |
Systems Methods Circuits and Associated Computer Executable Code
for Digital Catalog Augmentation
Abstract
Disclosed are methods, circuits, devices, systems and
functionally associated computer executable code for digital
catalog augmentation. A digital catalog interface module reads from
a digital catalog data storage, directly or indirectly, one or more
catalog data records constituting an offer listing within a digital
catalog, wherein the offer listing may include a description of a
specific product or service offering and/or links to execute a
transaction relating to the offering. The system includes a Review
Criteria and Sentiment Extractor (RCSE) to identify and convert one
or more reviews posted on a review forum into one or more data
records used to augment the offer listing within the digital
catalog.
Inventors: |
Aswani; Niraj; (Gujarat,
IN) ; Oza; Nilay; (Espoo, FI) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Klevu Oy |
Espoo |
|
FI |
|
|
Family ID: |
59275750 |
Appl. No.: |
15/399972 |
Filed: |
January 6, 2017 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62275252 |
Jan 6, 2016 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06F 40/253 20200101;
G06Q 30/0282 20130101; G06Q 30/0603 20130101; G06F 40/30 20200101;
G06F 40/232 20200101 |
International
Class: |
G06Q 30/06 20060101
G06Q030/06; G06F 17/28 20060101 G06F017/28; G06Q 30/02 20060101
G06Q030/02; G06F 17/27 20060101 G06F017/27 |
Claims
1. A digital catalog augmentation system comprising: a digital
catalog interface module to read from a digital catalog data
storage, directly or indirectly, one or more catalog data records
constituting an offer listing within a digital catalog, wherein the
offer listing includes a description of a specific product or
service offering; and a Review Criteria and Sentiment Extractor
(RCSE) to identify and convert one or more reviews posted on a
review forum into one or more data records used to augment the
offer listing within the digital catalog.
2. The digital catalog augmentation system according to claim 1,
wherein said catalog interface is further adapted to write to the
digital catalog data storage, directly or indirectly, one or more
catalog data records used to augment the offer listing within the
digital catalog.
3. The digital catalog augmentation system according to claim 1,
wherein said RCSE comprises a scrapper to scan through one or more
review forums and to identify and copy text from one or more
reviews relating to the offer listing in the digital catalog.
4. The digital catalog augmentation system according to claim 1,
wherein said scrapper is adapted to scrape reviews posted to a
review forum integral or otherwise associated with a digital
commerce platform of the digital catalog.
5. The digital catalog augmentation system according to claim 1,
wherein said scrapper is adapted to scrape reviews posted to a
review forum integral or otherwise associated with another digital
platform such as an online blog, a reviews website, a social
network, or any other server accessible through the internet.
6. The digital catalog augmentation system according to claim 3,
wherein said RCSE includes a natural language processor to extract
from the copied text at least one offering related criteria
associated with the offer listed in the digital catalog.
7. The digital catalog augmentation system according to claim 6,
wherein said natural language processor is further adapted to
extract from the copied text at least one assessment of or
sentiment towards the listed offering within a context of an
extracted criteria.
8. The digital catalog augmentation system according to claim 6,
wherein said natural language processor is context aware, such that
said processor cross-correlates pre-stored feature or attribute
information of an offering which is a subject of a specific review,
either a product or a service, as part of processing the review for
criteria and sentiment extraction about the offering.
9. The digital catalog augmentation system according to claim 1,
wherein augmenting the offer listing includes: (a) adding one or
more data records or files to be rendered as part of the offer
listing in the digital catalog; (b) editing one or more data
records or files to be rendered as part of the offer listing in the
digital catalog; and (c) removing one or more data records or files
to be rendered as part of the offer listing in the digital
catalog.
10. The digital catalog augmentation system according to claim 9,
wherein an added or modifier record includes at least one extracted
offering related criteria and an extracted assessment or sentiment
corresponding to the extracted criteria.
11. The digital catalog augmentation system according to claim 9,
wherein an added or modifier record expands a feature matrix
generated as part of the digital catalog offer listing.
12. The digital catalog augmentation system according to claim 1,
wherein extraction of a threshold number of reviews with negative
sentiment triggers one or more of: (a) generation of a report; (b)
a change in placement of the offer listing within the digital
catalog; (c) a change in catalog search engine result placement;
and (d) a suspension or delisting of the offering from the
catalog.
13. The digital catalog augmentation system according to claim 1,
wherein said review extractor further comprises a reviewer
assessment module adapted to assess credibility of a poster of one
or more offering reviews.
14. The digital catalog augmentation system according to claim 1,
wherein the review extractor further comprises review normalization
logic to correct transliterated (mixed-code) language by: (a)
referencing a knowledge base in various languages; or (b)
converting transliterated review language into original scripts
utilizing domain specific spell correction to obtain the nearest
possible words.
15. The digital catalog augmentation system according to claim 1,
wherein detection of a review about an offering being posted with
negative sentiment towards the offering triggers an automated
response to the review writer with a replacement offer and or a
monetary compensation offer.
16. The digital catalog augmentation system according to claim 1,
wherein augmenting the offer listing within the digital catalog
includes providing links to an alternate offering, wherein the
alternate offering is selected: (a) when one or more reviews
indicate superior properties or attributes of the alternate
offering; (b) when one or more reviews indicate dissatisfaction
with one or more properties, features or attributes of the listed
offering and the alternate has a higher rating corresponding to the
one or more properties, features or attributes; (c) when one or
more reviews indicate an unmet expectation with regard to one or
more properties, features, or attributes of the listed offering and
the alternate offering is known to meet the expectation with regard
to the one or more properties, features or attributes.
Description
RELATED APPLICATIONS
[0001] The present application claims priority from U.S.
Provisional Patent Application No. 62/275,252 filed on Jan. 6, 2016
and titled: `Systems Methods Circuits and Associated Computer
Executable Code for Opinion Mining on Consumer Reviews and
Applications thereof`. The full disclosure of the 62/275,252
Provisional Patent Application is hereby incorporated by reference
in its entirety.
FIELD OF THE INVENTION
[0002] The present invention generally relates to the fields of
online content publishing. More specifically, the present invention
relates to methods, circuits, devices, systems and functionally
associated computer executable code for augmenting digital
catalogs, for example, augmenting online catalogs using data mining
and automated analysis of offering reviews.
BACKGROUND
[0003] Opinions on products left on e-commerce websites and social
media sites have hidden golden dust for merchants. Often consumer
reviews are available in large quantity. Whilst it is difficult for
a merchant to read all of the reviews available, limiting the
number to a few can result into a biased consumer view.
[0004] Sentiment analysis is useful to understand overall mood in
the market for specific things. Typically the sentiment
summarizers, or feature based recommenders, work by maintaining a
set of features, pre identified from the product descriptions. In
such systems, vectors are used for maintaining counts of likes and
dislikes, often with time-series data. Such systems fail to
accommodate any new features other than those pre identified. This
is where opinion mining is helpful.
[0005] Some of the existing solutions use keywords to search and
analyze the results, wherein consumers may provide their aspired
preferences (e.g. specifying symptoms, compatibility criteria and
subjective hints; requesting trendy products and those that are
better than user specified brands). It is still beyond most
merchant's capability to: collect such information in today's fast
pace and large volume e-commerce environments, analyze the
information, update product catalogs based on the analysis and
provide a sophisticated search system that understands consumer's
queries.
[0006] Accordingly, there remains a need, in the field of online
content publishing and e-commerce, for solutions that may process
unstructured text, such as customer reviews/comments, to identify:
new features beyond those pre identified from the product
descriptions, the sentiment of customers (e.g. appraisal
expressions) towards the newly identified features of the product,
the general likes and dislikes of consumers and consumer groups,
strengths and weaknesses of products and services being sold,
consumers` wish lists, other competitors' strengths and weaknesses
and the like. Such insights can help merchants identify the right
`calls for actions`, for example, by allowing merchants to
automatically enrich the descriptions of their offerings within
digital catalogs and/or to report, remove, or present alternatives,
to offerings triggering negative reviews and sentiment.
SUMMARY OF THE INVENTION
[0007] The present invention includes methods, circuits, devices,
systems and functionally associated computer executable code for
augmenting a digital catalog, which catalog may comprise network
accessible data representing one or more product or service
offerings (herein after referred to as offering or offerings). A
digital catalog according to embodiments may include and/or
otherwise be associated with an e-commerce transaction system which
allows a catalog viewer to purchase an offering listed in the
catalog. A digital catalog according to embodiments may also
include, or be otherwise functionally associated with, one or more
online review zones where purchasers, consumers or users of a
listed offering may submit for posting a review providing an
(preferably personal) assessment of one or more characteristics or
attributes of the offering. According to some embodiments of the
present invention, there may be provided a digital catalog
augmentation system which may modify digital catalog data relating
to a specific digital catalog offering as a function of an
automated analysis of one or more reviews of the specific
offering.
[0008] A review scrapper integral or otherwise functionally
associated with the digital catalog augmentation system may access
and/or otherwise read reviews of one or more offerings listed in a
digital catalog. The scrapper may read reviews posted on the
digital catalog review zone, on a review zone linked to or
otherwise associated with the digital catalog, and/or to any
Internet accessible server where reviews of products or services
offered on the digital catalog may be posted.
[0009] According to some embodiments, review zones, Internet server
posted reviews and/or other sources of reviews linked to, or
otherwise associated with, the digital catalog may further include:
(1) any third party services allowing merchants, retailers,
producers, or distributers, to collect reviews, comments, opinions
and/or feedback; (2) social media sites; and/or (3) chat scripts
exchanged between users and/or between users and agents.
[0010] An automated review analyzer integral or otherwise
functionally associated with the digital catalog augmentation
system may include: (1) a Review Text Normalizer for normalizing
and correcting text of consumer reviews; and/or (2) a Review
Criteria and Sentiment Extractor for identifying, extracting and
characterizing: (a) the source, target and/or features of consumer
reviews/comments, and/or (b) new criteria, relating to offering(s)
in the digital catalog, and the sentiment expressed towards the new
criteria within the review/comment.
[0011] According to some embodiments, the System's Review Text
Normalizer may: (1) correct spelling mistakes within the text of
consumer reviews; (2) correct grammar mistakes within the text of
consumer reviews; and/or (3) handle mixed code language within
consumer reviews, in which multiple languages are used within the
same text section (e.g. same sentence, same paragraph), by
correcting transliterated language at least partially based on
domain specific spell correction.
[0012] According to some embodiments, the System's Review Criteria
and Sentiment Extractor may identify, within the normalized and
corrected text, new criteria by which to characterize an offering
within a digital catalog, wherein new criteria may include: (1)
features not mentioned in the catalog; (2) suitability and/or use
cases not mentioned in catalog; (3) compatibility and/or use
combinations, with other products, not mentioned in the catalog;
and/or (4) possible outcomes of usage, etc. The system's Review
Criteria and Sentiment Extractor may identify and/or extract the
sentiment expressed, within the text of the customer
review/comment, towards the identified new criteria.
[0013] According to some embodiments, the system's one or more
Feedback Modules and/or Applications may infer and utilize
knowledge from reviews and product descriptions for any combination
of the following actions:
[0014] (1) Removing, suspending and/or deprioritizing the showing
or presentation of an offering within the digital catalog; wherein
removed, suspended and/or deprioritized catalog offerings may
include offerings that received a certain number of harsh/negative
reviews and/or offerings that received reviews including a negative
sentiment beyond a predefined threshold of negativity level.
[0015] (2) Augmenting the digital catalog with data records
including new criteria and sentiment relating to offerings therein,
as expressed within one or more positive reviews. According to some
embodiments, augmenting a digital catalog may refer to any
combination of the following actions: (a) modifying digital catalog
records to be more correct, attractive and/or to include up to date
specific information about the product/service of the digital
catalog offering(s); (b) replacing digital catalog records with
more current and/or up to date records, wherein more current and/or
up to date information may relate to product/service specific
information/specifications; and/or (c) appending additional
information to existing digital catalog records and/or appending
new records.
[0016] (3) Identifying, within a review with a negative sentiment,
the specific targeted feature(s) of the offering towards which the
negative sentiment was expressed and augmenting the digital
catalog, by auto-responding to the review with a negative sentiment
with a listing/presentation of other versions of the offering (e.g.
alternative versions of products), wherein the listed/presented
other versions of the offering may include variations, or different
options, to the specific feature(s) of the offering that were the
target of the negative sentiment.
[0017] And/or (4) generating reports, or descriptions/updates,
relating to reviews including negative sentiment(s); wherein
generated reports may be relayed to a retailer and/or a producer of
the digital catalog offering's product or service which is the
subject of the review, thus providing the retailer/producer with
details of the negative sentiment and/or the criteria and possibly
product/service features to which it relates, allowing him to
correct or improve specific aspects of the product or service of
the offering. Negative sentiment review
reports/descriptions/updates may, for example, include: (a) details
of the corresponding catalog offering; (b) details of the
product/service features which were the target of the review;
and/or (c) new criteria addressed and identified in the review and
sentiment thereof.
[0018] According to some embodiments, the system's one or more
Feedback Modules and/or Applications, as part of, or in parallel
to, augmenting the digital catalog, may perform enrichment on the
extracted review/comment data to be augmented into the digital
catalog offering(s) records. Terms, titles and/or descriptions
extracted from a review/comment may be substituted with
corresponding terms, expressions, synonyms, parallel terms and/or
semantic normalizations. For example: the word `durable` may be
substituted with the expression `long lasting`; the term `child`
may be substituted with word `kid`; the title `a 13 years old boy`
may be substituted with `teenager`; and or the term `wife` may be
substituted, or broadened, to `female`(gender), `woman` and
`adult`. Substitute terms or descriptions may be registered to a
system data store in addition and association to their respective
terms, wherein some or all of the registered substitute terms may
be augmented into corresponding digital catalog offering(s)
records. The registered and/or augmented substitute terms may be
selected at least partially based on an estimation of their
positive effect on a potential customer to the corresponding
offering(s).
[0019] According to some embodiments, the system may assess the
credibility and/or trustworthiness of specific digital catalog
offerings reviewers and may allocate weights to specific reviews'
criteria and sentiments based thereof. Credibility and/or
trustworthiness assessment of a reviewer may be based on any
combination of the following factors: (1) Whether the reviewer is
estimated to be an end-user of the reviewed offering's product or
service or whether the review is based on a 3.sup.rd party/person
testimony, wherein end-user reviews may be allocated a higher
weight; (2) Whether the reviewer is estimated to be an expert in
the field of the offering's product or service, or whether he is a
private/unprofessional user, wherein expert reviews may be
allocated a higher weight; (3) Whether the reviewer, based on his
history of reviews, tends to focused on a specific domain(s) of
digital catalog offerings (e.g. mobile communication devices), or
whether his history of reviews relates to a substantially wide
range of offerings' fields/domains, wherein reviews made by
reviewers having a more domain/field focused reviews history may be
allocated a higher weight; and/or (4) Whether the reviewer, based
on his history of reviews or based on a specific review made, tends
to provide, or provided a specific, concise review(s), wherein more
concise reviews and/or reviews of more concise reviewers may be
allocated a higher weight.
[0020] According to some embodiments, customer reviews/comments,
whether positive or negative, may be filtered out from
consideration. Based on the analysis of the content of the
review/comment and/or based on the assessed credibility or
trustworthiness of the reviewer which is the source of the review,
as described above. Filtered out reviews/comments may, for example,
include: (1) reviews/comments estimated to be fake; (2)
reviews/comments estimated to have been made for fun or as a joke;
and/or (3) reviews/comments estimated to have been made for
political reasons.
BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The subject matter regarded as the invention is particularly
pointed out and distinctly claimed in the concluding portion of the
specification. The invention, however, both as to organization and
method of operation, together with objects, features, and
advantages thereof, may best be understood by reference to the
following detailed description when read with the accompanying
drawings:
[0022] FIG. 1A, is a block diagram of an exemplary system for
Automated Digital Catalog Augmentation in accordance with some
embodiments of the present invention;
[0023] FIG. 1B is a block diagram of the exemplary system for
Automated Digital Catalog Augmentation of FIG. 1A, further
comprising an offering(s) related sentiment reporting-logic
unit/module for compiling and relaying offering related sentiment
reports, in accordance with some embodiments of the present
invention;
[0024] FIG. 1C is a flowchart of the steps executed by the system
of FIGS. 1A and 1B as part of an exemplary process for Automated
Digital Catalog Augmentation, in accordance with some embodiments
of the present invention;
[0025] FIG. 2A shows, in greater detail, an exemplary:
review/comment scrapper, review text normalizer and review criteria
and sentiment extractor, in accordance with some embodiments of the
present invention;
[0026] FIG. 2B is a flowchart, showing the steps executed as part
of an exemplary process for review text normalization, in
accordance with some embodiments of the present invention;
[0027] FIG. 2C is a flowchart, showing the steps executed as part
of an exemplary process for review criteria and sentiment
extraction, in accordance with some embodiments of the present
invention;
[0028] FIG. 2D is a flowchart, showing the steps executed as part
of an exemplary process for review source trustworthiness
estimation, in accordance with some embodiments of the present
invention;
[0029] FIG. 3A shows, in greater detail, an exemplary `catalog
augmentation and/or `offering related sentiment reporting logic`
unit/module, in accordance with some embodiments of the present
invention;
[0030] FIG. 3B is a flowchart, showing the steps executed as part
of an exemplary process for catalog augmentation and sentiment
reporting, in accordance with some embodiments of the present
invention;
[0031] FIG. 4A shows a digital catalog product description table,
for an exemplary multipurpose pair of bicycle, prior to a review
based augmentation/enrichment process, in accordance with some
embodiments of the present invention;
[0032] FIG. 4B shows a system scrapped review of the multipurpose
pair of bicycle, in accordance with some embodiments of the present
invention;
[0033] FIG. 4C shows exemplary criteria and sentiments, extracted
from the multipurpose pair of bicycle review, in accordance with
some embodiments of the present invention; and
[0034] FIG. 4D shows a digital catalog product description table,
for an exemplary multipurpose pair of bicycle, following to a
review based augmentation/enrichment process, in accordance with
some embodiments of the present invention.
[0035] It will be appreciated that for simplicity and clarity of
illustration, elements shown in the figures have not necessarily
been drawn to scale. For example, the dimensions of some of the
elements may be exaggerated relative to other elements for
clarity.
DETAILED DESCRIPTION
[0036] In the following detailed description, numerous specific
details are set forth in order to provide a thorough understanding
of some embodiments. However, it will be understood by persons of
ordinary skill in the art that some embodiments may be practiced
without these specific details. In other instances, well-known
methods, procedures, components, units and/or circuits have not
been described in detail so as not to obscure the discussion.
[0037] Unless specifically stated otherwise, as apparent from the
following discussions, it is appreciated that throughout the
specification discussions utilizing terms such as "processing",
"computing", "calculating", "determining", or the like, may refer
to the action and/or processes of a computer, computing system,
computerized mobile device, or similar electronic computing device,
that manipulate and/or transform data represented as physical, such
as electronic, quantities within the computing system's registers
and/or memories into other data similarly represented as physical
quantities within the computing system's memories, registers or
other such information storage, transmission or display
devices.
[0038] In addition, throughout the specification discussions
utilizing terms such as "storing", "hosting", "caching", "saving",
or the like, may refer to the action and/or processes of `writing`
and `keeping` digital information on a computer or computing
system, or similar electronic computing device, and may be
interchangeably used. The term "plurality" may be used throughout
the specification to describe two or more components, devices,
elements, parameters and the like.
[0039] Some embodiments of the invention, for example, may take the
form of an entirely hardware embodiment, an entirely software
embodiment, or an embodiment including both hardware and software
elements. Some embodiments may be implemented in software, which
includes but is not limited to firmware, resident software,
microcode, or the like.
[0040] Furthermore, some embodiments of the invention may take the
form of a computer program product accessible from a
computer-usable or computer-readable medium providing program code
for use by or in connection with a computer or any instruction
execution system. For example, a computer-usable or
computer-readable medium may be or may include any apparatus that
can contain, store, communicate, propagate, or transport the
program for use by or in connection with the instruction execution
system, apparatus, or device, for example a computerized device
running a web-browser.
[0041] In some embodiments, the medium may be an electronic,
magnetic, optical, electromagnetic, infrared, or semiconductor
system (or apparatus or device) or a propagation medium. Some
demonstrative examples of a computer-readable medium may include a
semiconductor or solid state memory, magnetic tape, a removable
computer diskette, a random access memory (RAM), a read-only memory
(ROM), a rigid magnetic disk, and an optical disk. Some
demonstrative examples of optical disks include compact disk-read
only memory (CD-ROM), compact disk-read/write (CD-R/W), and
DVD.
[0042] In some embodiments, a data processing system suitable for
storing and/or executing program code may include at least one
processor coupled directly or indirectly to memory elements, for
example, through a system bus. The memory elements may include, for
example, local memory employed during actual execution of the
program code, bulk storage, and cache memories which may provide
temporary storage of at least some program code in order to reduce
the number of times code must be retrieved from bulk storage during
execution. The memory elements may, for example, at least partially
include memory/registration elements on the user device itself.
[0043] In some embodiments, input/output or I/O devices (including
but not limited to keyboards, displays, pointing devices, etc.) may
be coupled to the system either directly or through intervening I/O
controllers. In some embodiments, network adapters may be coupled
to the system to enable the data processing system to become
coupled to other data processing systems or remote printers or
storage devices, for example, through intervening private or public
networks. In some embodiments, modems, cable modems and Ethernet
cards are demonstrative examples of types of network adapters.
Other suitable components may be used.
[0044] Functions, operations, components and/or features described
herein with reference to one or more embodiments, may be combined
with, or may be utilized in combination with, one or more other
functions, operations, components and/or features described herein
with reference to one or more other embodiments, or vice versa.
[0045] The present invention generally relates to the augmentation
of a digital catalog based on customer or consumer reviews.
Throughout the specification, discussions utilizing terms such as
"review(s)", "comment(s)", "feedback", "opinion", or the like, may
refer to any form of input(s), received or retrieved, relating to
any characteristics or properties of a digital catalog offering(s).
The source of such reviews may be: a customer, a purchaser, a
consumer, a user, a reviewer, an automatic system or robot and/or
any other digital catalog offering(s) associated subject or
machine--or group thereof.
[0046] The present invention includes systems, methods, circuits,
and associated computer executable code for opinion mining on
consumer reviews and applications thereof. According to some
embodiments of the present invention, a system for opinion mining
on consumer reviews and applications thereof may comprise: (1) a
Review Text Normalizer for normalizing and correcting text of
consumer reviews; (2) a Consumer Review Processor for extracting
and characterizing the source, target and/or features of consumer
reviews and opinions; and/or (3) one or more Feedback Modules
and/or Applications for inferring and utilizing knowledge from
reviews and product descriptions.
[0047] According to some embodiments, a Review Text Normalizer may
comprise:
(a) a Consumer Review Spelling Corrector for ensuring product
domain specificity and higher accuracy in recognizing misspelled
features, wherein spellchecking is based on limited indexing of
domain specific dictionaries and product catalogs. (b) a Consumer
Review Grammar Corrector for handling grammatical inconsistencies
in consumer reviews by utilizing shallow parsing at the phrase
level, for example: "blue color I like not", "I do not like its
blue color", wherein as long as certain words of interest appear
within a close proximity words may be shuffled around their
respective head words, to convey meaning and opinion. And/or (c) a
Transliterated Language Corrector for dealing with mixed code
language in consumer reviews, by referencing knowledge base in
various languages, converting transliterated reviews, or sections
thereof, into original scripts, utilizing domain specific spell
correction to obtain the nearest possible words and relaxed
grammatical rules to figure out the associated opinion of
consumers.
[0048] According to some embodiments of the present invention, a
Consumer Review Processor may comprise:
(a) A Reviewer Analysis Classification and Characterization Module
for processing consumer reviews to extract source and estimate if a
reviewer is trustworthy/useful source, his review's weight, and/or
whether ho is an expert and/or a genuine consumer. According to
some embodiments a Reviewer Analysis Classification and
Characterization Module for processing consumer reviews may
implement one or more of the following methods and techniques,
and/or any combination thereof, for reviewer/review analysis and
characterization: [0049] According to some embodiments, considering
a given reviewer a trustworthy source may at least partially be
established based on the level of similarity of the given
reviewer's review, and the recommendation(s) it includes, to what a
majority of other reviewers say and recommend for the majority of
the features of the reviewed product. Whether, and/or to what
level, the review is in line with what others say may be
established by calculating sentiment score (described hereinafter
and in Appendix A) for every feature mentioned in the given
reviewer's review and comparing it against the overall `other
reviews averages` of the individual features. [0050] According to
some embodiments, considering a given reviewer a trustworthy source
may at least partially be established based on whether his reviews
are endorsed by other reviewers/consumers, or based on the number
of endorsements his reviews receive. [0051] According to some
embodiments, considering a given frequent reviewer a trustworthy
source may at least partially be established based on analysis of
his posts across different reviews, assessment of whether his posts
lead to specific domains, and, if his posts do lead to specific
domains, the reviewer may be considered a trustworthy source for at
least some of those specific domain(s). Reviews on/in other
topics/domains by the same reviewer may be treated as any other
`regular` consumer review. [0052] According to some embodiments,
considering a given reviewer a trustworthy source may at least
partially be established based on analysis of the reviewer's
`author name` and possibly of additional information about the
reviewer (e.g. organization name, designation, location) available
in the review metadata, and comparing the analyzed data with
information available in public Linked Open Datasets (LODs) such
as, but not limited to, DBPedia, Yago, Freebase etc. If the review
was collected from a website/web-location other than the e-commerce
website selling the product, and/or if the identity of the reviewer
can be established (e.g. it is listed in one of the (LODs)), than
such a person/entity may be considered a trustworthy source. In the
absence of a successful reviewer identification, the review may
receive the same weights as any other `regular` review on the
website. [0053] According to some embodiments, considering a given
reviewer a trustworthy source, and/or the usefulness of a given
review, may at least partially be established based on whether the
review was posted by an end-user of the product, wherein reviews
made by a reviewer estimated to be an end-user are considered more
useful and/or trustworthy. Various factors may be taken into
consideration for establishing whether a given review was posted by
an end-user or not, for example: (i) the presence of first person
pronouns (e.g. I, We, My, Our) and verbs indicating experience of
using a product (e.g. I tried, We found, In our experience, We
used); (ii) the use of non-technical vocabulary in the review, as
real consumers tend to focus on the use case of specific features
(e.g. commenting on clarity of picture rather than talking about
aperture when commenting on a camera); and/or (iii) the level of
the reviewer's distancing of himself from the use of
difficult/technical vocabulary, wherein a level of similarity (e.g.
a cosine similarity) between the text of the review and the product
description is calculated; assuming that the more technical the
review is the less likely it is that it was written by an end-user,
and based on the calculated similarity level, it may be established
whether the review was posted by a real consumer or a professional.
[0054] According to some embodiments, considering a given reviewer
a trustworthy source, and/or the usefulness of a given review, may
at least partially be established based on the length of the
review. As reviews tend to be short if genuine, longer reviews with
presence of second or third person pronouns (e.g. you, they) may
suggest that the reviews are posted by professional writers. (b) A
Pronominal Co-reference Resolver for identifying whether an opinion
is about the product overall or specific features by identifying
pronouns and the main subjects which the pronouns are associated
with. If no association can be established between the found
pronouns and targets, the product itself may be considered as a
target. Phrases without any target in the same sentence or in the
nearby context may be automatically associated with the main
product itself. (c) A Product Parts and Features Identifier for
identifying parts or features of products being discussed in
consumer reviews. Consumer comment on a specific attribute without
mentioning the actual value, the value may be obtained from the
product's specifications. For example, "its color is . . . ", "I
don't like its material", "size could have been bit . . . " etc.
Here, values of the mentioned attributes may be obtained from the
product's specifications, for example by cross correlating, or
identifying, the color/material/size attribute in the consumer
review/comment to the item (e.g. product part) he is viewing or
examining. (d) An Idiom Replacer for handling idioms in consumer
reviews, wherein based on a dictionary of idioms in different
languages, along with their corresponding sentiments as in the
positive, negative and neutral form, the dictionary may be used as
a source for looking them up in reviews, and optionally replacing
them with a matching actual meaning. (e) A Suitability Phrase
Identifier for identifying phrases revealing suitability aspects in
consumer reviews, wherein a gazetteer containing predefined sets of
words and patterns known to be commonly used for explaining the
same may be referenced and utilized for identification, for
example, the pattern: ((<ADVERB>)?<ADJ>):opinion
(<Noun>)?:category (<Preposition>):prep
(<Word>)+:suitableFor; that may match the following phrases:
"blazer is perfect for a business meeting": (perfect):opinion
(for):prep (a business meeting):suitableFor; and "the toy was
perfect for my 2 years old": (perfect):opinion (for):prep (my 2
years old): suitableFor. (f) An Opinion Normalizer for aggregating
opinions from consumer reviews and normalizing the attitudes (i.e.
appreciations or critics) of the consumer, possibly in addition to
normalizing the underlying features (i.e. targets), wherein
normalizing the attitudes may include enriching them with root
forms of the words and contextually relevant synonyms and semantic
categories. And/or, (g) a Sentiment Polarity Calculator for
obtaining sentiment polarity in consumer reviews, wherein based on
the product cluster to of a given product, and utilizing a
comparison API, other similar products in the cluster may be
fetched and their reviews aggregated, and the collective text may
provide context that is useful to determine the orientation of
opinion.
[0055] According to some embodiments of the present invention,
knowledge inferred from Consumer Review and Product Descriptions
may be utilized for updating and/or providing feedback to various
e-commerce systems and/or users thereof, exemplary applications may
include:
(a) Product Ontology Updating based on Consumer Review Opinions,
wherein, regular expressions over annotated reviews may be utilized
for identifying unknown use cases of products, thus enabling the
inferring of new knowledge about the corresponding reviewed
products. For Example: the consumer language "its color is
perfectly suitable for a business meeting" is cross correlated with
the respective product's specifications (e.g. a Jacket), inferring
that the following statement may be added to the product's ontology
(knowledge base): "light gray color is preferred by professionals
for business meetings". (b) Inferring of Dynamic Facets from
Consumer Reviews, wherein rule based information engineering
methods are utilized to process product descriptions and consumer
reviews, while searching for facets not explicitly highlighted by
the merchants of the product. For example: in the consumer language
"a perfect gift for infants"--the appraisal expression (i.e. a
perfect gift) is identified, and using the rules based information
engineering, as described hereinbefore, convert the concerning
"noun(s)" into a facet (e.g. gift), and present the facet to
consumers/customers searching for products. (c) Identifying
Products with Out-dated Features, wherein features represented as
numerical values are tracked and normalized for comparison purposes
(e.g. Camera 7MP>5MP), newer models succeeding the previous ones
(e.g. iPhone 6 succeeding iPhone 5) are tracked, and phrases added
to catalogs for such features are compared with the newer products
of the same category and the knowledge base is modified
accordingly. (d) Updating Product Catalogs Using Consumer Reviews,
wherein phrases with positive sentiments in the reviews are
identified and then `injected` back into the product catalog
allowing searches to be performed on newly discovered features. An
exemplary overall system of processing consumer reviews to update
product catalogs may execute some or all of the following steps,
and/or any combination thereof:
[0056] normalizing the text of consumer reviews to get rid of spell
errors;
[0057] using a microblog friendly POS tagger to annotate tokens
with grammatical tags
[0058] using shallow parsing techniques to find out sources and
appraisal expressions that express attitudes and targets;
[0059] assigning appropriate weights to the reviews by identifying
if a review was written by end-users, professionals, manufacturers
or spammers;
[0060] assigning to the respective features, a calculated polarity
of attitudes including those represented by idioms;
[0061] recognizing phrases used for explaining suitability aspects,
based on which new facets and suitability expressions are added to
the product catalogs; and
[0062] readjusting the ranking of products with out-dated
features.
And/or (e) Real-time Analysis of Consumer Reviews to Avoid Negative
Reviews, wherein text of reviews being written by consumers may be
monitored in real-time, consumer's dislikes (i.e. the target and
the attitude) are identified in real-time (i.e. prior to the
posting of the review) within negative consumer reviews as they are
being written or edited, the feature(s) of the product (e.g. color,
size) to which the complaint is targeted are identified, and
substantially similar products, not including the complained about
features, or including variations of these features, are presented
and offered to the consumer/customer.
[0063] In FIG. 1A, there is shown, in accordance with some
embodiments, a block diagram of an exemplary system for Automated
Digital Catalog Augmentation. In the figure, the Automated Digital
Catalog Augmentation System is shown to be functionally connected
to a Digital Commerce Platform. The Digital Commerce Platform
includes a digital/online catalog application server(s) for
receiving, optionally through the shown network gateway(s),
customer and point of sale (POS) orders for offerings included in
the digital catalog of the Digital Commerce Platform. The
digital/online catalog application server(s) is functionally
connected to a transaction and order fulfillment server(s) for
executing the received customer and point of sale (POS) orders and
for accordingly updating the digital catalog. The digital/online
catalog application server(s) is shown to be functionally connected
to the following data stores: a digital catalog code data store for
storing computer executable code for the generation, application
and/or management of the digital catalog; a digital catalog
offering details data store for storing details and descriptions of
product and/or service offerings included in the digital catalog;
and/or a digital catalog offerings related reviews/comments data
store for storing feedbacks from customers, consumers and/or other
users of product/service offerings in the digital catalog.
[0064] The digital/online catalog application server(s) shown, may
further receive, optionally through the shown network gateway(s),
consumer/customer/other reviews/comments for offerings included in
the digital catalog of the Digital Commerce Platform. The
digital/online catalog application server(s) may store the received
reviews/comments to the digital catalog offering(s) related
reviews/comments data store.
[0065] Customer and point of sale (POS) orders, and/or
consumer/customer/other reviews/comments, may be communicated to
the servers over a closed network or a direct communication session
(e.g. a POS network/direct/VPN connection), and/or from
computerized communication devices, over the Internet and through
the network gateway(s) shown.
[0066] The computerized communication devices shown in the figure
are provided as an example. Various computerized communication
devices/systems/components may be utilized to relay/upload digital
catalog offerings' orders and/or reviews/comments to the
digital/online catalog application server(s). Such
devices/systems/components may include, but are not limited to:
computing platforms, personal computers, laptops, tablets,
smartphones, smartwatches or other wearable devices, Internet
robots, smart house devices or appliances and/or any other digital
communication/networking able device/system. Furthermore, any of
the above listed devices/systems/components may be utilized to
relay/upload/post/share/endorse/like digital catalog offerings'
reviews/comments/opinions to product/service review web-site(s),
publications, blogs, social networks, immediate messaging platform
chats.
[0067] The Automated Digital Catalog Augmentation System shown,
includes a review/comment scrapper(s) for retrieving digital
catalog offerings reviews/comments from: the digital catalog
offerings reviews/comments data store of the Digital Commerce
Platform; and/or from one or more product/service review
web-site(s), publications, blogs, social networks and/or network
messages/chats, accessed through the shown network gateway(s) of
the Automated Digital Catalog Augmentation System.
[0068] The retrieved digital catalog offerings reviews/comments are
relayed to a review text normalizer for correcting grammar and
spelling errors in the text, normalizing the text and/or converting
transliterated text to original script/language based on knowledge
in the domain of respective product/service offerings in the
digital catalog.
[0069] Corrected and normalized reviews are processed by the shown
Review Criteria and Sentiment Extractor, utilizing a natural
language processor(s) including a criteria extractor and a
sentiment extractor, to identify within the corrected text of the
offerings reviews/comments, criteria relating to offering(s) in the
digital catalog and the sentiment expressed towards the new
criteria within the review/comment.
[0070] The shown catalog augmentation and/or offering related
sentiment reporting logic unit/module is utilized for estimating,
for each received review/comment, whether the review/comment is
positive. And, for each review/comment estimated to be positive:
(1) comparing the criteria and sentiment, identified within the
review/comment, to available digital catalog offering details
received through the shown offering details data reader of the
catalog interface, wherein the review/comment and the received
available digital catalog offering details relate to the
product/service of the same offering; (2) generating, for
identified review/comment criteria, not found (as part of the
comparison) within the received available digital catalog offering
details, data records including the new criteria and sentiment as
expressed within the review/comment; and (3) utilizing the offering
details data augmenter of the catalog interface to augment the
digital catalog, by updating the digital catalog offering details
data store of the Digital Commerce Platform with the generated data
records including the new criteria and sentiment, thus triggering
the addition of the new criteria and sentiment to
available/existing details/descriptions of offerings in the
catalog.
[0071] According to some embodiments, for received reviews/comments
estimated to be negative, the catalog augmentation and/or offering
related sentiment reporting logic unit/module may: (1) augment the
digital catalog with alternative products/services including
variations to specific features thereof, towards which the negative
review sentiment was expressed; and/or (2) remove from
presentation, or deprioritize the presentation (e.g. present as
later/last catalog offerings option) of, offerings which were the
target of the negative review/comment.
[0072] In FIG. 1B there is shown a block diagram of the exemplary
system for Automated Digital Catalog Augmentation of FIG. 1A,
further comprising an offering(s) related sentiment reporting-logic
unit/module for compiling and relaying offering related sentiment
reports, including details of negative sentiment expressed in a
review/comment and the criteria/feature of the product/offering
towards which it was expressed, to offering's point(s) of contact
(POC(s)) (e.g. retailer, producer, distributer).
[0073] In FIG. 1C there is shown, in accordance with some
embodiments, a flowchart of the steps executed by the system of
FIGS. 1A and 1B as part of an exemplary process for Automated
Digital Catalog Augmentation. The exemplary process shown includes
the following steps: (1) Accessing a digital catalog data store and
analyzing product or service offering(s) related data to compile a
set of offerings; (2) Scanning through one or more network/Internet
accessible servers/data-stores/data-repositories where offering(s)
related reviews/comments are posted; (3) Correlating specific
reviews/comments with corresponding offering(s) within the set
compiled from the catalog; (4) Using natural language processing
and/or artificial intelligence to extract one or more offering(s)
related criteria discussed, assessed, evaluated and/or otherwise
mentioned within one or more posted reviews/comments relating to
the specific offering(s); and/or (5) Using natural language
processing and/or artificial intelligence to extract sentiment
expressed in correspondence to the extracted criteria for the
specific offering(s).
[0074] Positive reviews/comments, including mostly or only positive
sentiment towards their respective extracted criteria, triggers:
(6) the Augmenting of digital offering(s) related data for the
specific offering(s) within the digital catalog (data store), using
extracted criteria and/or corresponding positive sentiment.
[0075] Negative reviews/comments, including mostly or only negative
sentiment towards their respective extracted criteria, triggers (7)
any combination of the following: (a) Notifying/Reporting to a
party responsible for the digital catalog, or offering(s) point of
contact, of extracted criteria with corresponding negative
sentiment; (b) Stopping or Deprioritizing the presentation, within
the catalog, of catalog offering(s) for which criteria with
corresponding negative sentiment was extracted; and/or (c)
Presenting alternative catalog offering(s) with alternatives to
specific criteria (e.g. product features related criteria) for
which corresponding negative sentiment was extracted.
[0076] In FIG. 2A there are shown in greater detail, in accordance
with some embodiments of the present invention: a review/comment
scrapper, a review text normalizer and a review criteria and
sentiment extractor; the operation of which may be described in
conjunction with the steps listed in the flowcharts of FIGS.
2B-2D.
[0077] The review/comment scrapper of FIG. 2A includes: a
web/network crawler, a data miner and/or a robot (bot), for finding
and retrieving reviews/comments related to digital catalog
offering(s) based on received details of digital catalog
offering(s) for which reviews/comments are to be scrapped. The
details of digital catalog offering(s) for which reviews/comments
are to be scrapped may be communicated, to the scrapper, by the
`catalog augmentation` and/or `offering related sentiment reporting
logic` unit/module (not shown) based on data from the offering(s)
details data reader of the catalog interface (not shown).
[0078] Reviews/comments are shown to be scrapped from the `digital
catalog offering(s) related reviews/comments data store` of the
digital commerce platform and/or from product/service review
websites, publications, blogs, social networks and/or chat/IM
messages.
[0079] The operation of the review text normalizer in FIG. 2A may
be described in conjunction with the steps listed in the flowchart
of FIG. 2B. The shown transliteration/mixed-code/multiple-language
normalizer initially scans scrapped reviews for transliterated,
mixed-code and/or multiple-language occurrences, and normalizes the
text by referencing knowledge base in various languages, codes,
slangs and/or domains, converting the found text occurrences into
an original script, in a single/unified code-type/language.
[0080] The shown spell corrector may correct the spelling of the
resulting normalized text, optionally utilizing domain specific
spellchecking rules, matching the domain of the product/service for
which reviews/comments are analyzed.
[0081] The shown grammar corrector corrects grammatical
inconsistencies in consumer reviews/comments, optionally utilizing
phrase level shallow parsing wherein words of interest in close
proximity are shuffled (e.g. pseudo randomly) to try and convey
meaning.
[0082] The operation of the review criteria and sentiment extractor
in FIG. 2A may be described in conjunction with the steps listed in
the flowchart of FIGS. 2C and 2D. The shown review source
(reviewer) extractor analyzes, classifies and characterizes the
sources (reviewers) of the corrected and normalized
reviews/comments, estimating if and to what level the source is
trustworthy/useful. The steps of the trustworthiness estimation
process executed by the review source trustworthiness estimator
shown in FIG. 2A, are listed and described in further detail in the
flowchart of FIG. 2D. the listed steps include: (1) Measuring the
similarity of a reviewer's review(s) to reviews made by other
reviewers, wherein higher similarity level indicates a more
trustworthy source; (2) Measuring the number of endorsements a
reviewer's review(s) received, wherein a higher number of
endorsements and/or more positive ones, indicate a more trustworthy
source; (3) Measuring the level of focus of a reviewer's reviews in
specific domain(s), wherein a higher domain focus level indicates a
more trustworthy source; (4) Checking the web/network place/origin
of a reviewer's review(s) and attempting to establish the
reviewer's identity, wherein reviews from web/network places other
than the web/network place selling, or directly selling, the
product/service of the digital catalog offering and/or reviews from
reviewer(s) whose identity was successfully established, indicate a
more trustworthy source; and/or (5) Measuring the length of the
review, wherein shorter, or more concise, reviews indicate a more
trustworthy source.
[0083] The review weight allocation logic shown in FIG. 2A,
allocates weights to reviews/comments and/or to criteria/sentiment
extracted therefrom, based on an aggregated weight of corresponding
reviewer's trustworthiness levels--as indicated by any combination
of outcomes of the above described reviewer assessment steps.
[0084] Returning now to FIG. 2A, there are shown a criteria
extractor and a sentiment extractor, the operation of the criteria
and sentiment extractors of FIG. 2A may be described in conjunction
with the following steps listed in the flowchart of FIG. 2C. the
listed steps include: (1) Identifying whether opinions/sentiments
within reviews target the overall product/service of the offering,
or specific criteria/features thereof; (2) Identifying the specific
parts, criteria, components and/or features of the reviewed
offering(s) targeted by the review; (3) Identifying idioms within
reviews and referencing an idiom dictionary (e.g. digital/web
dictionary/repository) to retrieve corresponding sentiment for
identified idioms, optionally replacing identified idioms with
corresponding `regular` language expressions; (4) Identifying
phrases revealing suitability aspects within reviews, by
referencing a dictionary/repository of predefined word sets
patterns commonly used for explaining suitability issues; (5)
Aggregating opinions from multiple reviews and normalizing the
sentiments/attitudes therein, optionally enriching the
sentiment/attitude expressions with root forms (e.g. root form
tags) of words and/or semantic categories thereof; and/or (6)
utilizing a comparison API/logic for finding
offerings/products/services belonging to, or associated with, the
same types/clusters of offerings/products/services of those
reviewed, retrieving and analyzing reviews of similar types, or
similarly clustered, offerings/products/services, to provide
further context/criteria and opinion attitude/sentiment for
currently analyzed offering(s) reviews.
[0085] In FIG. 3A there is shown in greater detail, in accordance
with some embodiments of the present invention, a `catalog
augmentation and/or `offering related sentiment reporting logic`
unit/module; the operation of which may be described in conjunction
with the steps listed in the flowchart of FIG. 3B.
[0086] The `catalog augmentation and/or `offering related sentiment
reporting logic` unit/module of FIG. 3A relays specific offering(s)
related data, received from the catalog interface reader, to the
system's review/comment scrapper described hereinbefore. In return,
the `catalog offerings available criteria/details` to `review(s)
extracted criteria` comparison logic shown, receives criteria and
sentiment of corresponding scrapped reviews, extracted by the
review criteria and sentiment extractor described hereinbefore,
optionally along with respective review allocated weights.
[0087] The `catalog offerings available criteria/details` to
`review(s) extracted criteria` comparison logic compares catalog
available/existing offerings related details data to extracted
review criteria and sentiment(s), outputting new
criteria/sentiments not yet in the catalog offering details. The
reviews weight factoring logic adjusts the weight(s) of the new
criteria/sentiments based on the received review allocated weights.
The digital catalog records generator, generates new digital
catalog records, based on the review(s)/comment(s) extracted
criteria/sentiments determined to be new (not yet in catalog
offering) and the weights allocated thereto, and forwards them to
the offering(s) details data augmenter of the catalog interface,
for augmentation into existing digital catalog records.
[0088] The digital catalog records generator, may further generate,
based on negative offering review(s), new digital catalog records
to cause the digital catalog to stop, or deprioritize, the
presentation of the corresponding offering(s) in the digital
catalog; and/or to present additional offering(s) with substitutes
to the specific product/service/offering feature(s) towards which
the negative sentiment in the review(s) was expressed.
[0089] The negative sentiment report generator, receives details of
product/service offering(s) and/or criteria/features thereof, that
received negative sentiment within the review(s); compiles a
corresponding report listing at least the negative sentiment
expressed within the review and the respective offering(s), or
offering(s) criteria/features, which are the target of the negative
sentiment; and relays the report to one or more point(s) of contact
associated with the product/service of the respective catalog
offering(s). Review(s) extracted new criteria may include: use
case, feature, usage combination, use outcome and/or outdated
feature, criteria types, as described hereinbefore.
[0090] In FIGS. 4A-4D there are respectively shown: a digital
catalog product description table, for an exemplary multipurpose
pair of bicycle, prior to a review based augmentation/enrichment
process (4A); an exemplary system scrapped review of the
multipurpose pair of bicycle (4B); exemplary criteria and
sentiments, extracted from the multipurpose pair of bicycle review
(4C); and the digital catalog product description table, of the
exemplary multipurpose pair of bicycle, following to a review based
augmentation/enrichment process (4D).
[0091] The digital catalog product description table of FIG. 4A
includes: digital catalog already-available offering descriptive
text, the product criteria it relates to and the sentiment of the
criteria. In the FIG. 4B review shown, extracted criteria and
sentiments are highlighted over the entire text of the review. FIG.
4C lists for each of the review extracted criteria: the criteria
type, the sentiment of the criteria, and the respective digital
catalog augmentation, or negative sentiment reporting, action it
triggers. In FIG. 4D the catalog product description table of FIG.
4A is shown following to the catalog augmentation based on FIG. 4C
criteria and sentiment.
[0092] The present invention may include a digital catalog
augmentation system with a digital catalog interface module to read
from a digital catalog data storage, directly or indirectly, one or
more catalog data records constituting an offer listing within a
digital catalog, wherein the offer listing may include a
description of a specific product or service offering and/or links
to execute a transaction relating to the offering. The system may
also include a Review Criteria and Sentiment Extractor (RCSE) to
identify and convert one or more reviews posted on a review forum
into one or more data records used to augment the offer listing
within the digital catalog. The catalog interface may further be
adapted to write to the digital catalog data storage, directly or
indirectly, one or more catalog data records used to augment the
offer listing within the digital catalog.
[0093] An offering listing within a digital catalog may be
generated by rendering one or more data records and/or data files
within a portion of the digital catalog. Offering listing is a
description, optionally with pictures of the offering, a cost of
the offering and/or instructions for purchasing the offering. The
digital catalog may part of and/or generated by a digital commerce
platform, retail or online.
[0094] According to embodiments, augmenting the offer listing
within a digital catalog, such as an online catalog, may include:
(a) adding one or more data records or files to be rendered as part
of the offer listing in the digital catalog; (b) editing one or
more data records or files to be rendered as part of the offer
listing in the digital catalog; and (c) removing one or more data
records or files to be rendered as part of the offer listing in the
digital catalog. An added or modifier record may include at least
one extracted offering related criteria and an extracted assessment
or sentiment corresponding to the extracted criteria. An added or
modifier record may expand a feature matrix generated as part of
the digital catalog offer listing.
[0095] An RCSE according to embodiments may include a scrapper to
scan through one or more review forums and to identify and copy
text from one or more reviews relating to the offer listing in the
digital catalog. The scrapper may be adapted to scrape reviews
posted to a review forum integral or otherwise associated with a
digital commerce platform of the digital catalog. The scrapper may
be adapted to scrape reviews posted to a review forum integral or
otherwise associated with another digital platform such as an
online blog, a reviews website, a social network, or any other
server accessible through the internet.
[0096] An RCSE according to embodiment may include a natural
language processor to extract from the copied text at least one
offering related criteria associated with the offer listed in the
digital catalog. The natural language processor may further be
adapted to extract from the copied text at least one assessment of
or sentiment towards the listed offering within a context of an
extracted criteria.
[0097] As part of understanding and extracting information
(offering related criteria, assessment, sentiment, etc.) from a
review of an offering, the natural language processor and or
another module integral or otherwise functionally associated with
an RCSE, according to embodiments, may apply normalization logic to
correct transliterated (mixed-code) language in a review, by: (a)
referencing a knowledge base in various languages; or (b)
converting transliterated review language into original scripts
utilizing domain specific spell correction to obtain the nearest
possible words. A natural language processor, or another module
integral or otherwise functionally associated with an RCSE,
according to embodiments, may be context aware, such that said
processor cross-correlates pre-stored feature or attribute
information of an offering which is a subject of a specific review,
either a product or a service, as part of processing the review for
criteria and sentiment extraction about the offering.
[0098] A digital catalog augmentation system according to
embodiments may, upon extraction of a threshold number of reviews
with negative sentiment towards a listed offering, may trigger one
or more of: (a) generation of a report; (b) a change in placement
of the offer listing within the digital catalog; (c) a change in
catalog search engine result placement; and (d) a suspension or
delisting of the offering from the catalog. According to further
embodiments, detection of a review about an offering being posted
with negative sentiment towards an offering may trigger an
automated response to the review writer with a replacement offer
and or a monetary compensation offer.
[0099] A digital catalog augmentation system according to
embodiments may augment an offer listing within a digital catalog
by providing links to an alternate offering, wherein the alternate
offering may be selected: (a) when one or more reviews indicate
superior properties or attributes of the alternate offering; (b)
when one or more reviews indicate dissatisfaction with one or more
properties, features or attributes of the listed offering and the
alternate has a higher rating corresponding to the one or more
properties, features or attributes; (c) when one or more reviews
indicate an unmet expectation with regard to one or more
properties, features, or attributes of the listed offering and the
alternate offering is known to meet the expectation with regard to
the one or more properties, features or attributes.
[0100] According to some embodiments, the digital catalog
augmentation system may include a reviewer assessment module to
assess a credibility of a poster of one or more offering
reviews.
[0101] The subject matter described above is provided by way of
illustration only and should not be constructed as limiting. While
certain features of the invention have been illustrated and
described herein, many modifications, substitutions, changes, and
equivalents will now occur to those skilled in the art. It is,
therefore, to be understood that the appended claims are intended
to cover all such modifications and changes as fall within the true
spirit of the invention.
* * * * *