Predicting-Depression-from-Social-Netwroking-Data-using-Machine-Learning-Techniques

This is a co-curricular Machine Learning project.

Aim: To detect depression from the social networking data, with as much accuracy as possible.

In this project we did the following things:

• Data Scraping from different social forums.
• Text Processing.
  • Extracting Basic Features
  • Count stopwords
  • Feature extraction
  • Lexical Diversity
  • Cleaning Data
• Natural Language Processing techniques.
  • Frequent word removal.
  • Word Cloud.
  • N-grams.
  • POS Tagging.
  • Sentiment Analysis.
  • Topic Modelling.
  • TF-IDF
• Comparing performances of different prediction models.
  • Support Vector Classifier (SVC)
  • Multinomial Naive Bayes (MultinomialNB)
  • K- Neighbors Classifier (KNN)

Best results were shown by Multinomial Naive Bayes model, 99.69%.

We also wrote a research paper for this project, which got accepted by the IEEE ICAC3N - 21 Scopus Indexed Conference, held in Dec, 2021.

Link to Published research paper - https://ieeexplore.ieee.org/document/9725402