Instructor: Dr. H. Faili
Summer 2022
Sentiment Analysis is one of the basic tasks in the field of natural language processing, which aims to examine the emotional state of the text. In other words, Entask aims to automatically detect whether the comment text contains positive or negative emotions (or some neutral) by extracting meaningful information from the text.
To succeed in the task, the following steps are recommended:
- Data analysis
First, use the tools to achieve the analysis and understanding of the presented dataset
-
Data processing
-
Data mining
-
Feature Extraction
For one task, you must use n-gram and TF-IDF methods for extracting and engineering or engineering and engineering.
- Modeling
Using classic models taught during the course, such as SVM, Logistic Regression, Decision Tree, ... develop models for Entask. Each student should develop at least three different models for one purpose. If needed, the tools available in the Scikit-Learn library should be used to adjust the models of the desired models and mentioned in the report. It is also necessary for students to pay attention to the issue of preventing overfitting or underfitting of models and include their solutions to solve a problem by mentioning the solution in the report. In the event of a data subject unbalancing, the solution used to deal with it should also be stated.
- Analysis of results