Superconduct-ML-Project

"Machine learning project using Superconduct dataset from Kaggle"

Superconduct-ML-Project

"Machine learning project using Superconduct dataset from Kaggle"

Project Title

Superconduct Critical Temperature Prediction using Machine Learning

Dataset

This project uses the [Superconductivity Dataset] (https://www.kaggle.com/datasets/willianoliveiragibin/superconduct-dataset) from Kaggle, which contains features describing the properties of superconducting materials and their corresponding critical temperatures.

• File used: superconduct-train-new.csv (27MB, locally stored)
• Rows: 21,263
• Features: 81 features including critical_temp (target variable)

Objective

To build and evaluate machine learning models that can accurately predict the critical temperature of superconductors and chemical properties.

ML Models Used

• Linear Regression
• Random Forest
• XGBoost
• Support Vector Regression (SVR)

Tools & Libraries

• R Programming
• tidyverse, caret, randomForest, xgboost, e1071, corrplot, data.table
• RMarkdown for report generation

Evaluation Metrics

• Root Mean Squared Error (RMSE)
• Mean Absolute Error (MAE)
• R² Score

Visualizations

The project includes the following:

• Correlation heatmaps
• Distribution of critical temperatures
• Feature importance plots
• Predicted vs Actual comparison graphs

Files Included

File : (Description) |Superconduct-ML-Project.R : (Full R script for data analysis & modeling) |Supercondut-ML-Project.Rmd : (RMarkdown report with code, plots, and results) |Superconduct-ML-Project.pdf: (Final PDF report exported from RMarkdown) |superconduct-train-new.csv : (Kaggle dataset file (tracked locally))