Improve Precategorized Collection Retrieval by Using Supervised Term Weighting Schemes

Due to the growth of Internet, it has not only become the medium for getting information, it has also become a platform for communicating. Social Network Service (SNS) is one of the main platform where Internet users can communicate by distributing, sharing of information and knowledge. Chatting has become a popular communication medium for Internet users whereby users can communicate directly and privately with each other. However, due to the privacy of chat rooms or chatting mediums, the content of chat logs is not monitored and not filtered. Thus, easing cyber predators preying on their preys. Cyber groomers are one of cyber predators who prey on children or minors to satisfy their sexual desire. Workforce expertise that involve in intelligence gathering always deals with difficulty as the complexity of crime increases, human errors and time constraints. Hence, it is difficult to prevent undesired content, such as grooming conversation, in chat logs. An investigation on two term weighting schemes on two datasets are used to improve the content-based classification techniques. This study aims to improve the content-based classification accuracy on chat logs by comparing two term weighting schemes in classifying grooming contents. Two term weighting schemes namely Term Frequency – Inverse Document Frequency – Inverse Class Space Density Frequency (TF.IDF.ICSdF) and Fuzzy Rough Feature Selection (FRFS) are used as feature selection process in filtering chat logs. The performance of these techniques were examined via datasets, and the accuracy of their result was measured by Support Vector Machine (SVM). TF.IDF.ICSdF and FRFS are judged based on accuracy, precision, recall and F score measurement.

Download Full-text

Ensemble of Classifiers and Term Weighting Schemes for Sentiment Analysis in Turkish

10.52460/src.2021.004 ◽

2021 ◽

Vol 1 (1) ◽

pp. 1-12

Author(s):

Aytuğ Onan ◽

Keyword(s):

Sentiment Analysis ◽

Language Processing ◽

Nearest Neighbor ◽

Text Messages ◽

Support Vector ◽

K Nearest Neighbor ◽

Term Weighting ◽

Text Documents ◽

Weighting Schemes ◽

Short Text

With the advancement of information and communication technology, social networking and microblogging sites have become a vital source of information. Individuals can express their opinions, grievances, feelings, and attitudes about a variety of topics. Through microblogging platforms, they can express their opinions on current events and products. Sentiment analysis is a significant area of research in natural language processing because it aims to define the orientation of the sentiment contained in source materials. Twitter is one of the most popular microblogging sites on the internet, with millions of users daily publishing over one hundred million text messages (referred to as tweets). Choosing an appropriate term representation scheme for short text messages is critical. Term weighting schemes are critical representation schemes for text documents in the vector space model. We present a comprehensive analysis of Turkish sentiment analysis using nine supervised and unsupervised term weighting schemes in this paper. The predictive efficiency of term weighting schemes is investigated using four supervised learning algorithms (Naive Bayes, support vector machines, the k-nearest neighbor algorithm, and logistic regression) and three ensemble learning methods (AdaBoost, Bagging, and Random Subspace). The empirical evidence suggests that supervised term weighting models can outperform unsupervised term weighting models.

Download Full-text