Tweets similarity classification based on Machine Learning Algorithms, TF-IDF and the Dynamic Case Based Reasoning
Mohamed Kouissi, El Mokhtar En-Naimi, Abdelhamid Zouhair · 2023
The research on the field of Twitter sentiment analysis, which aims to extract users’ sentiments through their public opinion about a given topic, has been increased and grown rapidly across a broad range of disciplines in the last decade. In this article, we propose a hybrid approach for Tweets similarity classification Based on Dynamic Case Based Reasoning approach, machine learning algorithms and Multi-Agent System. Our approach proposes a multi-agent adaptive system for Tweets similarity classification. It combines the Dynamic Case-Based Reasoning approach with the scientific measurement of keyword weight (Term Frequency- Inverse Document Frequency, TF-IDF). It consists of gathering and pre-processing tweets about a given topic and use a feature extraction to extract useful features. Machine Learning algorithms are then used for similarity content-based classification. Our approach is general and can be used to follow users’ tweets traces to predict their sentiments and provide them with an individualized content. In this study, Covid-19 tweets have been taken as an example.