This project implements an anime recommendation system using content-based and hybrid techniques. It processes anime data from a CSV file, conducts extensive text preprocessing on titles, themes, genres, and synopses, and utilizes TF-IDF vectorization for textual analysis.

Features:

• Content-Based Recommendation: Uses cosine similarity for anime suggestions based on textual features. Text preprocessing includes lowercasing, tokenization, special character removal, stop word removal, and lemmatization.
• Hybrid Recommendation: Combines content-based similarity and average ratings (65% content, 35% ratings) for personalized suggestions.

Data Preprocessing:

• Removes unnecessary columns.
• Standardizes text formats, removes spaces, special characters, and specific words/patterns from synopses.
• Numerical columns are processed to remove commas and scaled using MinMaxScaler.

Usage:

• Allows users to input specific anime titles for both content-based and hybrid recommendations.
• Provides recommendations for a predefined list of anime titles.

Libraries Used:

• Pandas: For data manipulation and preprocessing.
• NLTK: For tokenization, stop word removal, and lemmatization.
• Scikit-Learn: For TF-IDF vectorization, cosine similarity, and numerical scaling.