NLP-Driven Recommendation and Classification Algorithms
“Embarking on a comprehensive machine learning journey, this project reflects my end-to-end knowledge—from initial data collection to model training and deployment. With a keen focus on state-of-the-art technologies like Google’s Bidirectional Representation for Transformers (BERT), this undertaking exemplifies my commitment to staying ahead of the technological curve.”
Key Achievements
Holistic Project Execution: Demonstrated full-stack capabilities in machine learning, encompassing every phase from data procurement to model deployment.
Technological Acumen: Hands-on experience with industry-leading algorithms like BERT, fortifying the project’s classification and recommendation systems.
Core Competencies
Research Acumen, Effective Time & Task Management, Problem-Solving, Self-Driven Approach
Technical Proficiency
Machine Learning Algorithms:
Recommender Systems
Classification Systems
TF-IDF
BERT
Industry Tools & Software
Google Colab, Anaconda, StreamLit, GitHub
Project Specifics
Data Harvesting: Leveraged Beautiful Soup and Selenium to acquire targeted data from DiscourseHub Community forums.
Exploratory Data Analysis: Utilized TF-IDF for feature extraction and leveraged BERT to mitigate its limitations, thereby improving classification and recommendation capabilities.
Algorithmic Innovation: Implemented TF-IDF to obtain features in the form of vectorized matrix. Used BERT to overcome the shortcomings of TF-IDF in classifying an unlabelled post.
Deployment: Developed a user-friendly web application utilizing StreamLit..