Anime Recommender using Unsupervised Learning 🎯

Objective 🔍

Build a content-based anime recommendation system using clustering—no user history required.

---

About the Project 📘

Developed during a 4-week AI + Azure Internship (May 2025) by Edunet Foundation, this project recommends anime titles using unsupervised learning (K-Means) based on synopsis and genre.

---

Project Workflow 📊

A visual summary of data flow and tools used.

---

Dataset 📂

This dataset is used in this Python 3 project for analyzing anime rankings, ratings, genres, and other metadata.

Source:
🔗 https://www.kaggle.com/datasets/quanthan/top-15000-ranked-anime-dataset-update-to-32025

Modifications Made:

• Renamed the original file to anime_2025.csv for consistency with project naming.
• Removed the japanese_name column due to parsing issues.

Usage:

import pandas as pd
anime_df = pd.read_csv("anime_2025.csv")

The dataset includes 15,000 entries with key fields like:

• name, english_name, genres, score, type, episodes, source, rating, rank, popularity, etc.

Used for exploratory analysis, vectorization of synopses, and genre-based clustering.

---

Methodology 🛠️

1. Load Data: Anime info including synopsis, genres, and scores.
2. Preprocessing: Clean and vectorize synopses using TF-IDF.
3. Clustering: Apply K-Means to group similar anime.
4. Recommendation: Suggest titles from the same cluster as input anime.

---

Example 💡

Input: "Naruto"
Output: Similar anime like "Angel Beats!", "Platinum End" etc. 🎥

---

Run Instructions ▶️

1. Open Unsupervised_Anime_Recommendation_System.ipynb in Google Colab or locally.
2. Execute all cells:
   • Load dataset
   • Preprocess & vectorize
   • Fit model
   • Call the function below:

get_recommendations('Bleach') 
get_recommendations('Naruto')
get_recommendations('Bleech')  # typo test with fuzzy matching

---

Tech Stack 🧰

• Python (Colab)
• pandas, numpy, scikit-learn
• matplotlib, seaborn (optional for visuals)
• nltk (optional for NLP)

---

Future Scope 🔮

• Web interface (Flask/Streamlit)
• Add collaborative filtering
• Include user reviews
• Filter-based recommendations (genre/score)

---

Mentors & Inspiration 🙌

Inspired by Rohit Shelar Sir
Guided by:

• Rohit Shelar Sir
• Karthiga S Mam
• Thoufiq Ahmad Sir

---

Author 📌

Somapuram Uday
Final Year CST Student, GPREC(A), Kurnool, AP
Intern – Edunet Foundation | AI + Azure Virtual Internship 🎓

---

Next Steps ✅

• Fork & clone this repo
• Run the notebook
• Suggest or contribute improvements
• Reach out for feedback or collaboration! 💬