CLIAMS:
1. A method for providing feedback based on a topic based sentiment analysis of an input data, the method comprising steps of:
analyzing a contextual data corresponding to activities of one or more users from a plurality of data sources associated with a user device;
extracting a plurality of words corresponding to the activities of the one or more users from the contextual data;
identifying at least one of a topic of discussion from a predefined list by screening the plurality of words extracted from the contextual data;
detecting a sentiment corresponding to each word based on the topic detected from the contextual data corresponding to the activities of the one or more users;
calculating a sentiment score for each of the detected sentiment;
checking if a sentiment score of the contextual data is more than a predefined threshold; and
generating one or more feedback alerts for the user if the sentiment score is more than the predefined threshold.

2. The method of claim 1, wherein the one or more feedback alerts comprises at least one of providing a visual indication and an audio alert to the user for:
suggesting the user to correct a text having a negative tone;
changing a visual theme of a user interface of the user device; and
changing an audio theme;
wherein the visual indication comprises highlighting words or phrases using different colors, font, bold or smudging schemes.

3. The method of claim 1, wherein the sentiment is detected from the contextual data based on one or more granularities of a textual data, where the one or more granularities of the textual data comprises a word level, a phrase level and a sentence level.

4. The method of claim 3, wherein the method of detecting a sentiment from the contextual data at the word level comprises of:
categorizing each word extracted from the textual data into a Part-of-Speech (POS) category;
verifying whether each word extracted from the textual data is a named entity; and
classifying each word into at least one at least one of a positive sentiment, negative sentiment or neutral.

5. The method of claim 3, wherein the method of detecting a sentiment of the user from the contextual data at the word level further comprises:
analyzing each word in the textual information both syntactically and semantically;
verifying whether each word has a polarity value or not; and
classifying each word based on a tone of each word based on the polarity value into at least one category; wherein the sentiment categories classify the sentiments as positive, negative or neutral.

6. The method of claim 3, wherein the method of detecting a sentiment of the user from the contextual data at the sentence level comprises:
analyzing each sentence in the textual information based on syntactic clues and semantic clues; and
classifying each sentence based on a sentiment tone as at least one of positive sentiment, negative sentiment or neutral.

7. The method of claim 1, wherein the contextual data comprises of:
information on the activities of the user extracted from SMS, email and social media, where the activities of the user on the social media comprises at least one or more of tweets, replies or re-tweets to the tweets, posts, comments to other users' posts, opinions, feeds, connections, references, links to other websites or applications, or any other activities on the social media.

8. The method of claim 1, wherein providing feedback alerts to the user further comprises of:
extracting contextual data corresponding to activities of one or more users from a plurality of data sources;
extracting a plurality of words from the contextual data;
identifying the topic of discussion from a predefined list based on the extracted words;
detecting a sentiment associated with each word based on the identified topic along with a sentiment score;
creating a contextual vector based on the contextual words and sentiment score;
computing a dot product between the vector created and the pre-stored context word vectors;
selecting a plurality of syntactically, semantically and sentiment content-wise similar words based on one or more configurable parameters; and
displaying the plurality of similar words on the user device.

9. An apparatus for providing feedback based on topic based sentiment analysis on an input data, the apparatus comprising:
a data aggregator module adapted for:
analyzing a contextual data corresponding to activities of one or more users from a plurality of data sources associated with a user device; and
extracting a plurality of words corresponding to the activities of the one or more users from the contextual data;
a sentiment analysis engine, wherein the sentiment analysis engine comprises:
a topic identification module adapted for identifying at least one of a topic of discussion by screening the plurality of words extracted from the contextual data; and
a sentiment detection module adapted for detecting a sentiment corresponding to each word based on the topic identified from the contextual data corresponding to the activities of the one or more users;
a feedback response generator adapted for providing one or more feedback alerts to the user based on the sentiment associated with each word;
wherein the sentiment associated with each word is detected from the contextual data based on one or more granularities of a textual data, where the one or more granularities of textual data comprises a word level, a phrase level and a sentence level.

10. The apparatus of claim 9, wherein the sentiment analysis engine further comprises a sentiment ranking module adapted for:
calculating a sentiment score for each of the identified sentiment; and
checking if the sentiment score for each of the identified sentiment is more than a predefined threshold.

11. The apparatus of claim 9, wherein in detecting a sentiment from the contextual data at the word level, the sentiment analysis engine comprises a sentiment classifier module adapted for:
categorizing each word extracted from the contextual data into a Part-of-Speech (POS) category;
verifying whether each word extracted from the contextual data is a named entity; and
classifying each word based on the named entity status into at least one sentiment category , wherein the sentiment analysis engine classify the sentiments as at least one of positive, negative or neutral.

12. The apparatus of claim 9, wherein in identifying a sentiment from the contextual data at the word level, the sentiment analysis engine further comprises at least one module adapted for:
analyzing each word in the textual information both syntactically and semantically;
verifying whether each word has a polarity value or not; and
classifying each word based on a tone of each word based on the polarity value into at least one category, wherein the sentiment categories classify the sentiments as positive, negative or neutral.

13. The apparatus of claim 9, wherein in identifying a sentiment of the user from the textual data at the sentence level, the sentiment analysis engine comprises at least one module adapted for:
analyzing each sentence in the textual information based on syntactic clues and semantic clues; and
classifying each sentence based on the tone as at least one of positive, negative or neutral. ,TagSPECI:FIELD OF THE INVENTION

The present invention generally relates to sentiment analysis, and particularly relates to a method and apparatus for providing assistive feedback based on detected sentiment and topic identified from textual data.

BACKGROUND OF THE INVENTION

The widespread availability of consumer generated media (CGM), such as blogs, message boards, and comments on articles or social media network in general, has created a substantial new medium that allows ordinary consumers to voice their views about a product, service or idea. The existing technology employs sentiment analysis or opinion mining as automated tools to assess subjective information such as opinions, attitudes, and feelings expressed in text.

Current techniques often involve traditional keyword searching for particular negative or positive words such as hate, like, distaste, etc. to guesstimate the underlying sentiment of an article. Semantic-based approach generally relies on opinion work collection in the form of a sentiment dictionary or a large-scale knowledge base to assign sentiments to a text. For example, the opinion words refer to those sentiments possessing positive or negative sentiments such as ‘awesome,’ ‘great,’ ‘love’ or ‘terrible. However, this type of approach is not optimal because it may not appropriately capture the attitude or sentiment of the writer if the word is used in an unexpected way or in an unconventional sense.

Sentiment is generally measured as being positive, negative, or neutral. A common way to perform sentiment classification is to identify positive and negative words occurring in a text and use those words to calculate a score indicating the overall sentiment. A problem with this approach is that it does not account for the sentiment expressed by domain-specific words.

The foregoing examples of the related art and limitations related therewith are intended to be illustrative and not exclusive. Other limitations of the related art will become apparent upon a reading of the specification and a study of the drawings.

SUMMARY OF THE INVENTION

The primary objective of the present invention is to provide a method and apparatus for providing a context sensitive feedback for the user based on a detected sentiment and a related topic.

Another objective of the present invention is to provide a method and apparatus for providing sentiment analysis based on a textual information emanating from a user device.

Another objective of the present invention is to provide a method and apparatus for identifying a topic from the textual information extracted from a plurality of data sources.

Another objective of the present invention is to provide a method and apparatus for classifying the textual information based on sentiment and detected topic.

Another objective of the present invention is to provide a method and apparatus for assigning a score for the textual information based on the detected sentiment and topic.

The various embodiments of the present invention provide a method for providing feedback based on a topic based sentiment analysis of an input data. The method comprising of analyzing a contextual data corresponding to activities of one or more users from a plurality of data sources associated with a user device, extracting a plurality of words corresponding to the activities of the one or more users from the contextual data, identifying at least one of a topic of discussion from a predefined list by screening the plurality of words extracted from the contextual data, detecting a sentiment corresponding to each word based on the topic detected from the contextual data corresponding to the activities of the one or more users, calculating a sentiment score for each of the detected sentiment, checking if a sentiment score of the contextual data is more than a predefined threshold and generating one or more feedback alerts for the user if the sentiment score is more than the predefined threshold.

According to an embodiment of the present invention, the one or more feedback alerts comprises of providing one of a visual indication or an audio alert to the user to identify the text having a negative tone, suggesting the user to correct the text with the negative tone, change a visual theme of a user interface of the user device or change an audio theme of the user device. The visual indication comprises of highlighting words or phrases using different colors, font, bold or smudging schemes.

According to an embodiment of the present invention, the sentiment is detected from the contextual data based on one or more granularities of a textual data. Here the one or more granularities of the textual data comprise a word level, a phrase level and a sentence level.

According to an embodiment of the present invention, the method of detecting a sentiment from the contextual data at the word level comprises of categorizing each word extracted from the textual data into a Part-of-Speech (POS) category, verifying whether each word extracted from the textual data is a named entity and classifying each word into at least one sentiment category. The sentiment categories classify the sentiments as at least one of positive, negative or neutral.

According to an embodiment of the present invention, the method of detecting a sentiment of the user from the contextual data at the word level further comprises analyzing each word in the textual information both syntactically and semantically, verifying whether each word has a polarity value or not and classifying each word based on a tone of each word based on the polarity value into at least one category. The sentiment categories classify the sentiments as positive, negative or neutral.

According to an embodiment of the present invention, the method of detecting a sentiment of the user from the contextual data at the sentence level comprises analyzing each sentence in the textual information based on syntactic clues and semantic clues and classifying each sentence based on a sentiment tone as at least one of positive, negative or neutral.

According to an embodiment of the present invention, the contextual data comprises information on the activities of the user extracted from SMS, email and social media, where the activities of the user on the social media comprises at least one or more of tweets, replies or re-tweets to the tweets, posts, comments to other users' posts, opinions, feeds, connections, references, links to other websites or applications, or any other activities on the social media

Embodiments of the present invention further provide an apparatus for providing feedback based on topic based sentiment analysis on an input data. The apparatus comprises a data aggregator module, a sentiment analysis engine and a feedback response generator. The data aggregator module is adapted for analyzing a contextual data corresponding to activities of one or more users from a plurality of data sources associated with a user device and extracting a plurality of words corresponding to the activities of the one or more users from the contextual data. The sentiment analysis engine comprises a topic identification module adapted for identifying at least one of a topic of discussion by screening the plurality of words extracted from the contextual data and a sentiment detection module adapted for detecting a sentiment corresponding to each word based on the topic identified from the contextual data corresponding to the activities of the one or more users. The feedback response generator adapted for providing one or more feedback alerts to the user based on the sentiment associated with each word. The sentiment associated with each word is detected from the contextual data based on one or more granularities of a textual data. The one or more granularities of textual data comprise a word level, a phrase level and a sentence level.

According to an embodiment of the present invention, the sentiment analysis engine further comprises a sentiment ranking module adapted for calculating a sentiment score for each of the identified sentiment and checking if the a sentiment score for each of the identified sentiment is more than a predefined threshold.

According to an embodiment of the present invention, the sentiment analysis engine comprises a sentiment classifier module adapted for categorizing each word extracted from the contextual data into a Part-of-Speech (POS) category, verifying whether each word extracted from the contextual data is a named entity and classifying each word based on the named entity status into at least one sentiment category. The sentiment analysis engine classifies the sentiments as at least one of positive, negative or neutral. According to an embodiment of the present invention, the sentiment analysis engine further comprises at least one module adapted for analyzing each word in the textual information both syntactically and semantically, verifying whether each word has a polarity value or not and classifying each word based on a tone of each word based on the polarity value into at least one category. The sentiments herein are classified as positive, negative or neutral.

According to an embodiment of the present invention, the sentiment analysis engine further comprises at least one module adapted for analyzing each sentence in the textual information based on syntactic clues and semantic clues and classifying each sentence based on the tone as at least one of positive, negative or neutral.

The foregoing has outlined, in general, the various aspects of the invention and is to serve as an aid to better understanding the more complete detailed description which is to follow. In reference to such, there is to be a clear understanding that the present invention is not limited to the method or application of use described and illustrated herein. It is intended that any other advantages and objects of the present invention that become apparent or obvious from the detailed description or illustrations contained herein are within the scope of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The other objects, features and advantages will occur to those skilled in the art from the following description of the preferred embodiment and the accompanying drawings in which:

Figure 1 is a block diagram illustrating a feedback assistance apparatus with automated sentiment analysis and topic detection, according to an embodiment herein.

Figure 2 is a block diagram illustrating the functional modules of a sentiment analysis engine according to an embodiment herein.

Figure 3 is a flow chart illustrating a method of providing assistive feedback alerts based on a detected sentiment and topic, according to an example embodiment herein.

Figure 4 is a flow chart illustrating a method of providing assistive user feedbacks according to another example embodiment herein.

Figure 5 is a flow chart illustrating a method of providing assistive feedback according to another example embodiment herein.

Figure 6 is a block diagram of an exemplary feedback assistance apparatus, such as those shown in Figure 1, showing various components for implementing embodiments of the present subject matter.

Although specific features of the present invention are shown in some drawings and not in others. This is done for convenience only as each feature may be combined with any or all of the other features in accordance with the present invention.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a method and apparatus for providing feedback based on a topic based sentiment analysis of an input data. In the following detailed description of the embodiments of the invention, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is defined only by the appended claims.

Figure 1 is a block diagram illustrating a feedback assistance apparatus based on automated sentiment analysis and topic detection, according to an embodiment herein. The apparatus 100 herein comprises a data aggregator module 102, a sentiment analysis engine 104, a feedback response generator 106 and a display module 108. The apparatus 100 takes input from various data sources 101 associated with a user device. The various data sources comprise short messages 101a, emails 101b, or social media text 101c like status updates, broadcasts etc. The data aggregator module 102 analyzes the textual data based on the context corresponding to activities of one or more users from the data sources 101 and extracts a plurality of words from the contextual data.

The sentiment analysis engine 104 identifies sentiment of each word/sentence provided by the data aggregator module 104 along with the confidence score of the words. The sentiment analysis engine 104 also detects the topic of analysis. The topic comprises, but not limited to sports, movie, social life, politics, travel and the like. The sentiment analysis engine 104 also detects the sentiments different granularity of text such as at word level, phrase level and sentence level.

Further the feedback response generator 104 adapted creates one or more feedback alerts to the user based on the sentiment score associated with each word. The sentiment associated with each word is detected from the contextual data based on one or more granularities of a textual data such as word level, a phrase level and a sentence level. The feedback alerts are then displayed on a display module 105 of the user device.

Based on the application, the feedback response generator 104 generates the alerts or option message for the user to choose. The correction/modified/altered options/alerts are displayed to the user on a display module of the user device. If the user accepts the alerts or options from the feedback response generator, the change is applied else the next logical step is followed.

Figure 2 is a block diagram illustrating the functional components of a sentiment analysis engine according to an embodiment herein. The sentiment analysis engine 104 comprises a topic identification module 201, a sentiment detection module 202, a sentiment classifier module 203 and a sentiment ranking module 204. The topic identification module 201 identifies at least one of a topic of discussion by screening the plurality of words extracted from the contextual data. The sentiment detection module 202 detects a sentiment corresponding to each word based on the topic identified from the contextual data corresponding to the activities of the one or more users. The sentiment classifier module 203 categorizes each word extracted from the contextual data into a Part-of-Speech (POS) category. Further the sentiment classifier module 203 classifies each word into a sentiment category. The sentiments herein are classified as positive sentiment, negative sentiment or neutral sentiment.

The sentiment ranking module 204 then calculates a sentiment score for each of the identified sentiment and verifies if the sentiment score for each of the identified sentiment is more than a predefined threshold. If the sentiment score does not cross the threshold then next logical action is to be continued as specified.

Figure 3 is a flow chart illustrating a method of providing assistive feedback alerts based on a detected sentiment and topic, according to an embodiment herein. At step 301, the contextual data corresponding to activities of one or more users are analyzed. Here the contextual data comprises textual data extracted from various data sources and text converted voice data. Further a plurality of words is extracted from the contextual data at step 302. At step 303, identify a topic of discussion based on the words extracted from the contextual data. Here the topics are identified from a predefined list of topics stored in a data repository associated with a user device. Further detect a sentiment associated with each word based on the identified topic and calculate an associated sentiment score at step 304. At step 305, verify if the calculated sentiment score targeted at specific text granularity is greater than a preset threshold. If the sentiment score is greater than the threshold, then generate one or more feedback alerts for the user at step 306. If the sentiment score is less than the threshold, then perform a predefined action at step 307.

Figure 4 is a flow chart illustrating a method of providing feedback alerts according to an example embodiment of the present invention. At step 401, the contextual data corresponding to activities of one or more users is extracted from a plurality of data sources. Further extract a plurality of words from the contextual data at 402. At 403, identify the topic of discussion from a predefined list based on the extracted words. At 404, detect a sentiment associated with each word based on the identified topic along with a sentiment score. Further, classify each word based on the sentiments at 405. At 406, perform selective text user interface change based on polarity of words and word level sentiment.

The embodiments herein further enables selective text user interface change based on a word level sentiment. The embodiments herein provide the user with an option to highlight all words which have some kind of polarity. For example, based on the polarity, words will be marked in blue if it is positive and in red if it is negative. Through such distinctions and making an aggregate of the same, the user can comprehend the objective component of the text easily. Moreover the embodiments herein, dictates whether the user is chatting is in a high state of emotion depending on the amount of sentiment bearing words in the text.

Figure 5 is a flow chart illustrating a method of providing feedback alerts according to another example embodiment of the present invention. At step 501, the contextual data corresponding to activities of one or more users is extracted from a plurality of data sources. Further extract a plurality of words from the contextual data at 502. At 503, identify the topic of discussion from a predefined list based on the extracted words. At 504, detect a sentiment associated with each word based on the identified topic along with a sentiment score. Further, create contextual vector based on the contextual words and the sentiment score at 505. Then compute the dot product between the vector created and the pre-stored context word vectors at 506. At 507, select the plurality of words which are syntactically, semantically and sentiment content-wise similar and based on the configurable parameters. At 508, display the plurality of selected words on the user device.

Figure 6 is a block diagram of an exemplary feedback assistance apparatus 100, such as those shown in Figure 1, showing various components for implementing embodiments of the present subject matter. In Figure 6, the feedback assistance apparatus 600 includes the processor 601, the memory 602, a display module 105, an input device 604, and a cursor control 606, a read only memory (ROM) 608, and a bus 610.

The processor 601, as used herein, means any type of computational circuit, such as, but not limited to, a microprocessor, a microcontroller, a complex instruction set computing microprocessor, a reduced instruction set computing microprocessor, a very long instruction word microprocessor, an explicitly parallel instruction computing microprocessor, a graphics processor, a digital signal processor, or any other type of processing circuit. The processor 601 may also include embedded controllers, such as generic or programmable logic devices or arrays, application specific integrated circuits, single-chip computers, smart cards, and the like.

The memory 602 and the ROM 608 may be volatile memory and non-volatile memory. The memory 602 includes a data aggregation module 102, a sentiment analysis engine 103 and a feedback response generator module 104 for providing assistive feedback alerts based on a detected sentiment and topic according to one or more embodiments described above. A variety of computer-readable storage media may be stored in and accessed from the memory elements. Memory elements may include any suitable memory device(s) for storing data and machine-readable instructions, such as read only memory, random access memory, erasable programmable read only memory, electrically erasable programmable read only memory, hard drive, removable media drive for handling compact disks, digital video disks, diskettes, magnetic tape cartridges, memory cards, and the like.

Embodiments of the present subject matter may be implemented in conjunction with modules, including functions, procedures, data structures, and application programs, for performing tasks, or defining abstract data types or low-level hardware contexts. The data aggregation module 102, a sentiment analysis engine 103 and a feedback response generator module 104 may be stored in the form of machine-readable instructions on any of the above-mentioned storage media and may be executed by the processor 601. For example, a computer program may include machine-readable instructions, that when executed by the processor 601, cause the processor 601 to provide generate and provide feedback alerts to the user based on the identified sentiments and topic, according to the teachings and herein described embodiments of the present subject matter. In one embodiment, the computer program may be included on a compact disk-read only memory (CD-ROM) and loaded from the CD-ROM to a hard drive in the non-volatile memory.

The bus 610 acts as interconnect between various components of the feedback generator apparatus 600. The components such as the display module 105, the input device 604, and the cursor control 606 are well known to the person skilled in the art and hence the explanation is thereof omitted.

According to an example embodiment herein, consider a case wherein an assistive feedback application is installed in a first user device. The textual information entered by the first user is analyzed before the message is sent over to the second user. Based on the detected sentiment level, an alert/warning is provided to stop the first user from sending the text to the second user. This is done only at those instances when the sentiment of the text is overtly negative and crosses a predefined threshold.

The embodiments herein also enables provides for enabling assistive feedback for replacing words based on synonymy and sentiments, by using assistive feedback application as a thesaurus. Another use of sentiment based assistive feedback technology is to use it as thesaurus. Generally, thesaurus works based on the word synonymy i.e. they can be used to replace words which have the same meaning. However, if the sentiment contend is embedded in the thesaurus lookup, it is possible to get words which are similar and also which preserve the sentiment content of the text. For example, in one case, the editor gives option to replace the word “horrible” using a normal thesaurus, whereas in a second case the editor uses thesaurus which sorts the words options based on the sentiment contend and similarity of the word.

The present embodiments have been described with reference to specific example embodiments; it will be evident that various modifications and changes may be made to these embodiments without departing from the broader spirit and scope of the various embodiments. Furthermore, the various devices, modules, and the like described herein may be enabled and operated using hardware circuitry, for example, complementary metal oxide semiconductor based logic circuitry, firmware, software and/or any combination of hardware, firmware, and/or software embodied in a machine readable medium. For example, the various electrical structure and methods may be embodied using transistors, logic gates, and electrical circuits, such as application specific integrated circuit.

Although the embodiments herein are described with various specific embodiments, it will be obvious for a person skilled in the art to practice the invention with modifications. However, all such modifications are deemed to be within the scope of the claims. It is also to be understood that the following claims are intended to cover all of the generic and specific features of the embodiments described herein and all the statements of the scope of the embodiments which as a matter of language might be said to fall there between.