A framework for emotion mining from text in online social networks(final)
Transcript of A framework for emotion mining from text in online social networks(final)
![Page 1: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/1.jpg)
A Framework for Emotion Mining from Text in Online Social Networks
Mohamed Yassine Hazem HajjDepartment of Electrical and Computer Engineering
American University of BeirutBeirut, Lebanon
E-mail: {mmy18, hh63}@aub.edu.lb
Advisors: Yin-Fu Huang Student: Chen-Ting Huang
Published in:Data Mining Workshops (ICDMW), 2010 IEEE
International Conference
![Page 2: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/2.jpg)
Abstract & Introduction
A new framework is proposed for characterizing emotional interactions in social networks, and then using these characteristics to distinguish friends from acquaintances.
The framework includes a model for data collection, database schemas, data processing and data mining steps.
The paper presents a new perspective for studying friendship relations and emotions’ expression in online social networks
![Page 3: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/3.jpg)
Text mining techniques are performed on comments retrieved from a social network.
The purpose is not to identify specific emotions but rather to tell if the text contains emotions or not.
In other words, if the text is subjective reflecting the writer’s affect and emotional state or if it is factual and objective.
Abstract & Introduction
![Page 4: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/4.jpg)
Related Work
SentiWordNet: a publicly available lexical resource for opinion mining
NTUSD
Face Book
![Page 5: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/5.jpg)
Social Network – text example
Post
Comment
![Page 6: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/6.jpg)
Row data collection
Row Data Collection
![Page 7: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/7.jpg)
Posts Comments User Information Friends
Row Data Collection
The texts to be mined for emotions
![Page 8: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/8.jpg)
Lexicons Development
Lexicons Development
![Page 9: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/9.jpg)
This step deals with the informal language of social networka. Social acronyms
a. Emotions
Lexicons Development
![Page 10: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/10.jpg)
Lexicons Development
This step deals with the informal language of social networkc. Foreign Languages Lexicons: It is also necessary to take into consideration the use of foreign
languages.
![Page 11: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/11.jpg)
Feature generation
Feature generation
![Page 12: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/12.jpg)
Feature generation
Computes new features from available raw data collected in step 1 to assess subjectivity of text.
It uses word-matches to existing affective lexicons (by SetiWordNet) and employs new lexicons developed in step 2
Comments collected from step 1 along with the features computed in this step will be stored in the Sentiment Mining Database
![Page 13: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/13.jpg)
Feature generation
In order to assess the subjectivity of the text, several features need to be computed.
a. Affective lexicon based on SentiWordNet
b. Based on intentional misspelling errors and grammatical markers (such as punctuation and capitalized letters)
c. Based on social acronyms, interjections and emoticons.
![Page 14: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/14.jpg)
Data preprocessing
Data Preprocessing
![Page 15: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/15.jpg)
Data Preprocessing
Feature selection was performed based on a correlation measure to remove redundant attributes.
![Page 16: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/16.jpg)
Data Preprocessing
The goal was to deduce whether the text is subjective or not.
Example: knowing whether the writer uses few or many punctuation marks would definitely denote some sort of subjectivity.
Discretization was done using clustering. The k-means algorithm was performed on each attribute with k=3 or k=4.
The values in each attribute were normalized using Min-Max normalization which mapped them to the range [0;1].
![Page 17: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/17.jpg)
Creating a Training model for text subjectivity
Training Model for Text Subjectivity
![Page 18: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/18.jpg)
Training Model for Text Subjectivity
The training was done in step 5 using the k-means clustering algorithm with k =3.
a. Objective or factual texts,
b. Moderately subjective texts
c. Subjective texts
The output of the model was the centroids of the three clusters and is used in step 6 for classifying the comments with respect to subjectivity.
![Page 19: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/19.jpg)
Text subjectivity classification
Text subjectivity classification
![Page 20: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/20.jpg)
This step uses the centroids generated in the previous steps.
Employs the k-nearest neighbor algorithm with k=1 to classify all comments into one of three subjectivity levels.
The manual annotation was then compared to the predicted result by the model.
Text subjectivity classification
![Page 21: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/21.jpg)
Model Evaluation
eg: Comment 1• Repeated letters: ‘o’ ‘i’ ’l’ ’u’ ’.’ ’o’ ’h’• Emotion: ‘ :P ’ ‘ :) ’• Affective words:• Acronyms: mwahhh
![Page 22: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/22.jpg)
Model Evaluation
k-means algorithm & Min-Max normalization
![Page 23: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/23.jpg)
Model Evaluation
![Page 24: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/24.jpg)
Evaluating the model was done by applying the same above procedure on the sample of 850 comments.
Model Evaluation
81%
84%
![Page 25: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/25.jpg)
Friendship Classification
Friendship Classification
![Page 26: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/26.jpg)
Friendship Classification
We collected the comments shared for 1213 pairs of users.
We used the output of step 6 which is text subjectivity classification in order to get the subjectivity measure of all comments shared between the pair.
SVM algorithm was used to predict whether the pair consists of close friends or just acquaintances.
Every user reported a list of his/her close friends. Using 10-fold cross-validation, the SVM model reported an accuracy of 87%.
![Page 27: A framework for emotion mining from text in online social networks(final)](https://reader033.fdocuments.us/reader033/viewer/2022060122/559510b91a28ab00108b46d6/html5/thumbnails/27.jpg)
Conclusion It presents a new perspective for studying friendship relations and
emotions’ expression in online social networks
We developed new lexicons that cover common expressions used by online users.
The model predicted the right class with 88% accuracy. When the resulting model was used to predict relationship strength between two users, the prediction reported an accuracy of 87%.