A SMART HYBRID ENHANCED RECOMMENDATION AND PERSONALIZATION ALGORITHM USING MACHINE LEARNING - Project Presented for Master of Science in Computer Science - May 2024

Aug 5, 2026

Upload by Eliana

Streaming services make recommendation accuracy essential for user satisfaction. This project proposes the Smart Hybrid Enhanced Recommendation and Personalization Algorithm (SHERPA), a machine learning approach for movie suggestions. SHERPA integrates TF-IDF content-based filtering with Alternating Least Squares (ALS) plus Weighted Regularization to improve tailored ranking quality. Evaluation uses over 50 million Netflix ratings from roughly 480,000 users spanning 17,000 titles, comparing SHERPA against traditional hybrid models via RMSE during training, testing, and validation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages59

Reading Time2 h 29 min

AccessFree to download

ABSTRACT

ACKNOWLEDGEMENTS

LIST OF TABLES

LIST OF FIGURES

CHAPTER ONE: INTRODUCTON

Background

Significance

Purpose

CHAPTER TWO: LITERATURE SURVEY

Traditional Machine Learning Approaches

Modern Machine Learning Approaches

CHAPTER THREE: DATA PREPARATION

Data Collection

Movie Titles Dataset File Description

Movie Ratings Dataset File Description

Data Cleaning

Data Pre-processing

Data Parsing

Data Structuring

Format Handling Issues

CHAPTER FOUR: METHODOLOGY

Term Frequency-Inverse Document Frequency

Singular Value Decomposition

Mathematical Formulation of SVD

The Mechanics of SVD

SVD to Make Predictions

Alternating Least Squares

Update Procedure

Alternating Least Squares with Weighted Regularization

The Loss Function

Example Application of ALS

Content Based Filtering

Collaborative

A SMART HYBRID ENHANCED RECOMMENDATION AND PERSONALIZATION ALGORITHM USING MACHINE LEARNING A Project Presented to the Faculty of California State University, San Bernardino In Partial Fulfillment of the Requirements for the Degree Master of Science in Computer Science by Aswin Kumar Nalluri A SMART HYBRID ENHANCED RECOMMENDATION AND PERSONALIZATION ALGORITHM USING MACHINE LEARNING A Project Presented to the Faculty of California State University, San Bernardino by Aswin Kumar Nalluri May 2024 Approved by: Dr. Yan Zhang, Advisor, Computer Science and Engineering Dr. Yunfei Hou , Committee Member Dr. Jennifer Jin , Committee Member © 2024 Aswin Kumar Nalluri ABSTRACT In today’s age of streaming services, the effectiveness and precision of recommendation systems are crucial in improving user satisfaction. This project introduces the Smart Hybrid Enhanced Recommendation and Personalization Algorithm (SHERPA) a cutting-edge machine learning approach aimed at transforming how movie suggestions are made. By combining Term Frequency Inverse Document Frequency (TF-IDF) for content based filtering and Alternating Squares (ALS) with Weighted Regularization for filtering SHERPA offers a sophisticated method for delivering tailored recommendations. The algorithm underwent evaluation using a dataset that included over 50 million ratings from 480,000 Netflix users encompassing 17,000 movie titles. The performance of SHERPA was meticulously compared to traditional hybrid models demonstrating a 70% enhancement in prediction accuracy based on Root Mean Square Error (RMSE) metrics during training, testing and validation phases. These findings highlight SHERPAs capability to understand and cater to users’ subtle preferences representing an advancement in personalized recommendation systems. ACKNOWLEDGEMENTS I owe a heartfelt thanks to my advisor Dr. Yan Zhang for her wisdom, continuous support, and encouragement for my project "Smart Hybrid Enhanced Recommendation and Personalization Algorithm (SHERPA)" . I am sincerely thankful to Dr. Yunfei Hou and Dr. Jennifer Jin for agreeing to be part of my committee and for believing in me for the successful completion of my project. Your insights have helped me, and your willingness to share your expertise has deeply enriched my work. I am also thankful to High Performance Computing Program team at California State University San Bernardino for utilizing the HPC resources to finalize my results. A special note of gratitude to my family and friends, whose encouragement has been my source of strength and motivation throughout the project. TABLE OF CONTENTS ABSTRACT .......................................................................................................... iii ACKNOWLEDGEMENTS .....................................................................................iv LIST OF TABLES ............................................................................................... viii LIST OF FIGURES ............................................................................................... ix CHAPTER ONE: INTRODUCTON ....................................................................... 1 Background................................................................................................ 1 Significance ............................................................................................... 1 Purpose ..................................................................................................... 2 CHAPTER TWO: LITERATURE SURVEY ........................................................... 3 Traditional Machine Learning Approaches................................................. 3 Modern Machine Learning Approaches ..................................................... 3 CHAPTER THREE: DATA PREPARATION ......................................................... 5 Data Collection ....................

FAQs

What is SHERPA and what problem does it address?

SHERPA is a smart hybrid algorithm designed to improve movie recommendation quality in streaming services by learning users’ preferences more precisely.

How does SHERPA combine content-based and collaborative techniques?

SHERPA uses TF-IDF for content-based filtering and Alternating Least Squares (ALS) for collaborative filtering, enhanced with Weighted Regularization.

What dataset and evaluation metrics were used to test SHERPA?

The evaluation uses a dataset with over 50 million Netflix ratings from 480,000 users and 17,000 movie titles. Performance is compared using Root Mean Square Error (RMSE) across training, testing, and validation phases.

Aug 5, 2026

Upload by Eliana