MACHINE LEARNING JOURNEY REPOSITORY

Welcome to my Machine Learning Journey repository! This repository showcases a variety of machine learning projects, encompassing both supervised and unsupervised techniques. Here, I demonstrate end-to-end machine learning development, including data collection, manipulation, exploratory data analysis (EDA), and feature engineering. This repository serves as a testament to my commitment to mastering machine learning concepts and practical applications.

Overview

This repository contains implementations of a variety of machine learning algorithms, highlighting my proficiency in both foundational and advanced techniques. Each project illustrates a specific problem, methodology, and solution, complete with detailed code and analysis.

Key Projects

1. Titanic Survival Prediction

• Description: Predicting survival on the Titanic using passenger data.
• Algorithm: Naive Bayes
• Accuracy: 77%

2. Holiday Package Recommendation

• Description: Recommending holiday packages based on customer preferences.
• Algorithms and Accuracies:
  • Random Forest: 92%
  • AdaBoost: 62%
  • Gradient Boost: 95%
  • XGBoost: 93%

3. Car Price Prediction

• Description: Predicting car prices based on various features.
• Algorithms: Random Forest and KNN
• Evaluation (R² Scores):
  • Random Forest: 0.93
  • KNN: 0.88

4. Market Basket Analysis

• Description: Discovering associations in transaction data.
• Algorithms: Apriori and FP-Growth
• Application: Association rule mining for market basket insights.

5. Iris Species Classification

• Description: Classifying iris species based on flower measurements.
• Algorithms: Naive Bayes and Decision Tree
• Accuracy: 100%

6. Diabetes Progression Prediction

• Description: Predicting the progression of diabetes in patients.
• Algorithm: Naive Bayes

7. Housing Price Prediction

• Description: Predicting housing prices based on various economic indicators.
• Algorithm: Multiple Linear Regression
• Evaluation (R² Score) : :0.58

8. Forest Fire Weather Index Prediction

• Description: Predicting Forest FWI value from Algerian Forest Fire Datasets .
• Algorithm: Linear Regression, Lasso(L1 Regularization), Ridge(L1 Regularization) and ElasticNet
• Evaluation (R² Scores):
  • Linear Regression : 0.98
  • Lasso Regression : 0.94
  • Ridge Regression : 0.98
  • Elaticnet : 0.87

Implemented Algorithms

Supervised Learning

• Regression: Simple Linear Regression, Multiple Linear Regression, Polynomial Regression, Logistic Regression, Ridge Regression, Lasso Regression and ElasticNet
• Classification: Decision Trees, Random Forest, Gradient Boost, AdaBoost, XGBoost, Naive Bayes, K-Nearest Neighbors (KNN), Support Vector Machines (SVM)

Unsupervised Learning

• Dimensionality Reduction: Principal Component Analysis (PCA)
• Clustering: K-means, Hierarchical Clustering, DBSCAN
• Association Rule Mining: Apriori, FP-Growth

Getting Started

Each project folder contains a Jupyter notebook or Python script, detailing the data preprocessing steps, model training, evaluation, and results. Simply clone the repository and navigate to the respective project directory to explore the code.

git clone https://github.com/DeepakMishra99/Machine-Learning-Practice.git