Iris Project - Species classification with ML

🌸 Iris Project – Species Classification with Machine Learning

Project Overview:
This project focuses on the analysis and supervised classification of the well-known Fisher's Iris dataset, aiming to accurately predict the species of iris flowers (Iris setosa, Iris versicolor, and Iris virginica) based on numerical features: sepal length, sepal width, petal length, and petal width.

Objectives

• Perform exploratory data analysis (EDA) to identify key patterns and relationships between variables.

• Create interactive 3D visualizations to explore class separability.

• Build and evaluate machine learning classification models.

• Optimize model hyperparameters using cross-validation techniques.

Key Tasks Performed

• Exploratory Data Analysis:
  Investigated feature distributions, correlations, and class separability using pandas, seaborn, and matplotlib.

• Interactive 3D Visualization:
  Built a 3D scatter plot with plotly, mapping three features and coloring points by species to enhance interpretability.

• Supervised Modeling:
  Implemented and compared two classic classification algorithms using scikit-learn:

  • Logistic Regression

  • K-Nearest Neighbors (KNN)

• Model Evaluation:
  Assessed performance using metrics such as accuracy, precision, recall, F1-score, and confusion matrices.

• Hyperparameter Tuning:
  Applied GridSearchCV to optimize KNN parameters (e.g., number of neighbors, weighting scheme, distance metric) and improve generalization.

Results

• Both models achieved a 93.3% accuracy on the test set.

• Cross-validation suggested a marginal improvement for the optimized KNN, although this did not translate into significantly better test set performance.

• Iris setosa was classified with perfect accuracy; minor misclassifications occurred between versicolor and virginica, consistent with their natural overlap.