ML Classification Algorithms

Simple python implementation of various ML binary and multiclass classification algortithms.

Requirements

• python
• numpy
• pandas
• matplotlib
• mpl_toolkits

1. Logistic regression

code file: 1_logisitic_binary.py

data file: data3.xlsx

This dataset contains 4 features for different instances as four columns and fifth column is for class label.

Logistic regression algorithm for the binary classification using hold-out cross-validation technique with 60 % of instances as training and the remaining 40% as testing.

2. Multiclass logistic regression

code file: 2_logistic_multiclass.py

data file: data4.xlsx

This dataset contains 4 features for different instances as four columns and fifth column is for class label (3 classes).

Multiclass logistic regression algorithm using both “One VS All” and “One VS One” multiclass coding techniques. Hold-out cross-validation approach (60% training and 40% testing) used for the selection of training and test instances.

3. Multiclass logistic regression classifier using 5-fold cross-validation

code file: 3_logisitic_multiclass_k_fold.py

data file: data4.xlsx

Multiclass logistic regression classifier using 5-fold cross-validation approach.

4. Likelihood ratio test (LRT) for the binary classification

code file: 4_LRT_binary.py

data file: data3.xlsx

Likelihood ratio test (LRT) for the binary classification using hold-out cross-validation technique with 60 % of instances as training and the remaining 40% as testing.

5. Maximum a posteriori (MAP) decision rule for multiclass classification

code file: 5_MAP_multiclass.py

data file: data4.xlsx

Maximum a posteriori (MAP) decision rule for multiclass classification task using hold-out cross-validation approach (70% training and 30% testing).

6. Maximum likelihood (ML) decision rule for multiclass classification

code file: 6_ML_multiclass.py

data file: data4.xlsx

Maximum likelihood (ML) decision rule for multiclass classification using hold-out cross-validation approach (70% training and 30% testing).

License

Distributed under the MIT License. See LICENSE.txt for more information.