Welcome to my machine learning repository! Here, you'll find various projects and datasets that I have worked on over the past few months. Each folder in this repository corresponds to a specific project or dataset. Below, I've provided a brief overview of each folder along with the URL to access the repository.
This project explores various unsupervised learning techniques on the Fashion MNIST dataset The goal is to apply dimensionality reduction and clustering algorithms to identify patterns and group similar fashion items together.
In this project, I aimed to predict housing prices in various districts across California using several supervised regression algorithms. I utilized the California Housing Prices dataset sourced from Kaggle, which provides valuable information on district-specific attributes such as population, median income, and median housing price.
In this project, I aimed to predict forest cover types based on various attributes using several supervised classification algorithms. The Forest Cover Type Prediction dataset was used for this analysis. After performing Exploratory Data Analysis (EDA) to understand the dataset's structure and relationships, I employed various classification methods, including Linear Classification, SVM, KNN Classification, Decision Trees, Adaboost and MLP. Additionally, I created an Ensemble Classifier using the top 3 and top 5 performing models to improve predictive accuracy.
This project about the Forest Cover Type Prediction Backend API, which serves three supervised classification models for predicting forest cover types based on the Forest Cover Type Prediction dataset. The API provides endpoints to access the top classifier, 2nd best classifier, and the ensemble classifier.
Time Series Analysis is a statistical technique used to analyze and forecast data that is collected over time at regular intervals. It involves studying patterns, trends, and seasonality in time-dependent data to make predictions for future values. This documentation covers the basics of time series analysis, including Autocorrelation and Partial Autocorrelation, and explores three traditional forecasting algorithms: Autoregressive (AR), Moving Average (MA), Autoregressive Moving Average (ARMA), and Autoregressive Integrated Moving Average (ARIMA). These algorithms are applied to two datasets: monthly sunspots and daily minimum temperatures.
This project presents a comprehensive analysis of a multivariate multi-step time series dataset related to household energy consumption. The analysis focuses on applying two different approaches to forecast energy consumption: Autoregressive (AR) models and Long Short-Term Memory (LSTM) networks. The dataset includes multiple variables, enabling the exploration of complex relationships and patterns within energy consumption data.