Predicting Iron Content in Iron Ore Using Machine Learning

Overview

This project focuses on predicting the iron content in iron ore using machine learning techniques. The primary goal is to develop a model that can accurately estimate the iron content based on various features extracted from iron ore data. The project involves data preprocessing, feature engineering, model training, and evaluation.

Methodology

Data Preparation

1.Data Import and Exploration:

The dataset is imported from a CSV file containing raw data on iron ore samples. The data includes features like iron concentration, silica content, and other relevant variables.

2.Data Cleaning:

Missing values are handled, and outliers are identified and removed using the Z-score method to ensure data quality.

3.Feature Engineering:

New features are created by combining similar columns to reduce dimensionality and improve model performance. For example, air flow and level data from different flotation columns are averaged.

Model Development

1.Scaling Data:

MinMaxScaler is used to normalize the feature values, ensuring that all features contribute equally to the model.

2.Model Training:

Several regression algorithms are evaluated, including Linear Regression, Ridge Regression, Lasso Regression, K-Nearest Neighbors (KNN) Regressor, and Decision Tree Regressor. Pipelines are created for each algorithm to streamline the process of scaling and fitting the models.

3.Model Evaluation:

Cross-validation is performed to assess model performance, using metrics such as Root Mean Squared Error (RMSE) and R² score. The KNN Regressor is found to have the best performance based on RMSE and accuracy.

Results

The final model, KNN Regressor, is trained on the scaled data and evaluated on the test set. The performance metrics, including RMSE and R² score, are reported. A scatter plot visualizes the relationship between actual and predicted iron content, providing insights into model accuracy.

Usage

To run this project, clone the repository and follow these steps:

• Install the required dependencies from requirements.txt.
• Download the Dataset from link "https://www.kaggle.com/datasets/edumagalhaes/quality-prediction-in-a-mining-process".
• Execute the Jupyter notebook or Python script to preprocess the data, train the models, and evaluate their performance.