1 of 19

CAPSTONE PROJECT��Book Recommendation System

​

PRESENTED BY

STUDENT NAME:Achhuta Nand Jha

COLLEGE NAME:GLA University

DEPARTMENT:Computer Engineering & Applications

EMAIL ID:achhuta.jha_cs.da23@gla.ac.in

AICTE STUDENT ID:STU662a2083d78941714036867

2 of 19

OUTLINE

  • Problem Statement
  • Proposed System/Solution
  • System Development Approach  
  • Algorithm & Deployment  
  • Result (Output Image)
  • Conclusion
  • Future Scope
  • References

​

3 of 19

PROBLEM STATEMENT

With the ever-growing number of books being published each year, readers often struggle to find books that match their tastes and preferences. Traditional methods such as best-seller lists or general reviews lack personalization. Hence, there's a need for an intelligent system that helps users discover books based on their interests and reading history.

4 of 19

PROPOSED SOLUTION

The proposed Book Recommendation System leverages collaborative filtering and content-based filtering techniques to suggest books personalized to user preferences.

  • Data Collection: Used public book datasets including user ratings and metadata.�
  • Filtering Techniques:�
    • Collaborative Filtering: Based on user similarity and rating patterns.�
    • Content-Based Filtering: Based on features like author, title, genre, etc.�
  • User Interface: Simple UI to display top 50 books and get personalized recommendations.�
  • Deployment: Hosted using Flask on Render platform.�

​

5 of 19

SYSTEM  APPROACH

  • Technologies Used:�
    • Python, Pandas, NumPy�
    • Scikit-learn (for modeling)�
    • Flask (for backend)�
    • HTML/CSS (for frontend)�
    • Render (for deployment)�
  • Libraries:�
    • pandas, numpy, scikit-learn, flask, pickle, requests, etc.

6 of 19

ALGORITHM & DEPLOYMENT

Algorithm Used:

  • Collaborative Filtering: Based on cosine similarity between users or items.�
  • Content-Based Filtering: TF-IDF or metadata similarity.�

Data Input:

  • Book titles, user ratings, authors, genres, publication years.�

Training Process:

  • Precomputed similarity matrix based on user ratings and book features.�
  • Model serialized with pickle for quick deployment.

7 of 19

ALGORITHM & DEPLOYMENT

Prediction:

  • Given a selected book, return the top 3 most similar books.�

Deployment:

  • Web app deployed using Flask.�
  • Hosted on Render using a free tier service.

8 of 19

RESULT

9 of 19

RESULT

10 of 19

RESULT

11 of 19

RESULT

12 of 19

RESULT

13 of 19

RESULT

14 of 19

RESULT

15 of 19

RESULT

Please note: The website will take approximately 1 minute to load after clicking 'Show Project'. Kindly wait while the system initializes.

Github link

Show Project

​

​

Please note: The website will take approximately 1 minute to load after clicking 'Show Project'. Kindly wait while the system initializes.

16 of 19

CONCLUSION

The system provides personalized book recommendations using a hybrid filtering approach.�

Easy to use, responsive interface hosted online.�

Helps readers explore new titles aligned with their interests.�

Demonstrated effective performance with a clean UI and relevant suggestions.

​

​

17 of 19

FUTURE SCOPE

Incorporate user login and history tracking for dynamic recommendations.�

Use deep learning-based models like BERT for better understanding of book content.�

Enable multilingual recommendations.�

Expand the dataset for broader genre coverage and global reach.�

Add features like “Recently Trending”, “Critic’s Picks”, and “User Reviews”.

18 of 19

REFERENCES

Kaggle Datasets (Books, Ratings)��CampusX YouTube Tutorials and Resources�

Render Deployment Platform

​

GitHub Link: Link

​

​

19 of 19

Thank you