Homayoun S.'s Projects
In this project, I used the RAVDESS dataset with eight emotions, each at two intensities. I built a model extracting five features from speech signals to classify emotions, providing valuable insights.
In this project, I analyzed Coffee Bean Sales using Python, employing libraries such as pandas, matplotlib
an Agentic customer support system that automate responses to common queries, prioritize urgent issues, and escalate complex queries to human agents.
In this project, I analyzed flight delays using Python, and I leveraged libraries such as pandas, matplotlib, and seaborn.
In this project, I built a chatbot to answer customer simple inquiries about restaurants, such as displaying the food menu and contact, and providing information about the number of available tables for reservation.
Personal Portfolio
In this project I built a linear regression algorithm from scratch to predict the price of the houses based on different features
python-review-scraping-imdb-movie-site-with-BeautifulSoup
In this project, I have built an interactive dashboard for a supermarket dataset using streamlit framework
In this project I designed a knowledge graph focused on Napoleon's history. I built a RAG application using this data and improved the output of LLM using the relationship between nodes
Predict house price with linear regression
The objective of this project was to predicting housing prices for New York City. To achieve this goal, I performed web scraping to collect a dataset of 17,000 records from realtor.com, covering all cities in the state of New York. Employing transfer learning, I utilized a model trained on 10,000 records to further train a model using 2,000 records
In this repository, I implemented a RAG (Retrieval-Augmented Generation) framework using Faiss for efficient similarity search and integrated it with the T5 model within the LangChain framework.
In this project I built a web developer agent based on reflection agent concept. This agent can reflect on generated website template and optimize it by itself
For this project, I aimed to perform sentiment analysis on IMDB movie reviews. My dataset consisted of over 36,000 reviews, each accompanied by movie ratings ranging from 0 to 10. The primary objective was to construct a machine learning model capable of categorizing reviews into three sentiment classes: negative, neutral, and positive.