project1-LDA

README.md

@@ -1,8 +1,96 @@
-# Project 1 
+# README
+
+**TEAM:**
+- Rakesh Reddy - A20525389
+- Geeta Hade - A20580824 
+- Nishant Khandhar - A20581012
+- Amogh Vastrad - A20588808 
+
+
+
+## Overview
+
+
+This project implements a Linear Discriminant Analysis (LDA) model for dimensionality reduction and classification. The model projects high-dimensional data into a lower-dimensional space, maximizing the distance between classes. It is particularly useful for visualizing data with a reduced number of features.
+
+
+
+### When to Use the Model
+
+- **Classification Tasks**: Suitable for classifying data into different classes.
+- **Multivariate Normal Distribution Assumption**: Effective when data for each class follows a normal distribution.
+- **Dimensionality Reduction**: Reduces data to lower dimensions while maintaining class separability.
+
+
+## Running the Program
+
+### Step 1: Generate the Data
+
+To generate synthetic data, use the following command:
+ ```bash
+py .\testDataGenerator.py -N <Sample size> -f <features> -c <classes> -seed 10000 -output_file generated_data.csv
+ ```
+
+
+
+Example:
+ ```bash
+py .\testDataGenerator.py -N 150 -f 8 -c 5 -seed 10000 -output_file generated_data.csv
+
+ ```
+
+
+### Step 2: Run the Program
+
+Execute the final LDA model:
+Note:  Final_LDA.ipynb is our  final code and notebook  file to test our program,
+
+ ```bash
+Final_LDA.ipynb 
+
+ ```
+The program execution steps are shown in the notebook file
+ ```bash
+Final_LDA.ipynb 
+
+ ```
+
+
+
+
+## Model Testing
+
+### Step 1: Training on Available Dataset (Iris)
+
+- Initially, the model was built using the Iris dataset, a common classification problem involving classifying flowers into one of three different iris species.
+
+### Step 2: Generated Data
+
+- The model was tested using synthetic data generated from `testDataGenerator.py`, allowing for testing under various conditions with variable classes, samples, and features.
+
+### Step 3: Manual Train-Test Split
+
+- Data is split into training and test sets to evaluate performance, preventing overfitting and ensuring accurate training.
+
+### Step 4: Accuracy Calculation
+
+- The model’s predicted classes are compared against the true classes to calculate accuracy.
+
+## User-Exposed Parameters
+
+Users can adjust the following parameters to tune model performance:
+
+- **nComponents**: Number of components to keep in LDA, controlling dimensionality reduction.
+- **Regularization**: Adds regularization to the within-scatter matrix, preventing overfitting in high-dimensional data.
+- **Solver**: Allows choice of eigenvalue computation method, introducing flexibility.
+- **Test Size**: Determines the percentage of data allocated for testing.
+- **Random State**: Ensures reproducibility of train-test splits and generated data.
+
+## Known Limitations
+
+The implementation may encounter difficulties with:
+
+1. **Highly Imbalanced Data**: Struggles when class sizes are noticeably imbalanced.
+2. **High Dimensionality with Few Samples**: May exhibit irregular behavior when the number of features exceeds the number of samples. Dimensionality reduction using PCA can be a solution.
 
-Put your README here. Answer the following questions.
 
-* What does the model you have implemented do and when should it be used?
-* How did you test your model to determine if it is working reasonably correctly?
-* What parameters have you exposed to users of your implementation in order to tune performance? (Also perhaps provide some basic usage examples.)
-* Are there specific inputs that your implementation has trouble with? Given more time, could you work around these or is it fundamental?