Sitaras/Data-Mining

Project 1: 🎬🍿 Movie-Recommendation-System, Project 2: 📰🔍Fake News Detection System

Data-Mining

Project 1 - Movie Recommendations

The dataset of this project contains Netflix movies.

• Project Implementation
• Project Description

Part 1

In the first part of the project we observe the dataset and produce statistics about the content of the dataset. Some of the statistics are:

• Number of movies/series.
• Country with the most content.
• Year with the most content.
• The popularity of each genre for every country.

Part 2

We implement a recommendation system in order to recomend similar movies to a given movie. In order to represent each movie we tried the two following representations:

• Bag Of Words (BoW)
• TF-IDF

In order to compute the similarity between the repsesentations we used:

• Jaccard-Tanimoto coefficient
• Cosine Similarity

PS: if the notebook cannot be opened on github, you can view it via the Jupiter nbviewer:

• Visit: https://nbviewer.org/
• Paste the link of the notebook (https://github.com/giannhskp/Data-Mining/blob/main/Project1.ipynb)

Project 2 - Fake/True News Classification

• Project Implementation
• Project Description

Given a dataset with news articles we should train a model that classifies each article as fake or True. We try different ways to represent the text of each article, such as:

• Bag Of Words
• TF-IDF
• Word2Vec

Also, we use different models in order to compare their performance. The models that we used are:

• Logistic Regression
• Naive Bayes
• Support Vector Machines (SVM)
• Random Forest

Finaly we compare the performance between every combination of representation/model.