Decision tree is coded with Python and tested with iris dataset. The interface of the decision tree class is designed to be compatible with sk-learn.
Various linear regressions for multi variables and L1 and L2 regulizations are coded using python numpy vectors and matrices. The following figure shows a polynomial fitting without regulization.
The implementation of logistic regression, L1 and L2 regulizations and gradient descent using python numpy matrix and vector. Logistic regression is used to evaluate the probability of a binary outcome using input variables. For example, binary logistic regression can be used to classify a Donut dataset.
Comparison of various classifiers applied to the facial recognition data from the Kaggle facial expression recognition challenge https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge. Seven categories of facial expressions are classified: Angry, Disgust, Fear, Happy, Sad, Surprise, and Neutral.
Five category class of the E-commerce data is classified using logistic regression and various implementations of neural networks using Pure Numpy, Sklearn, or Tensorflow.
Implementation of Neural Network, L1 and L2 regulizations and gradient descent using Python numpy matrix.
Solving machine learning problems using the scikit learn package. Scikit-learn can dramatically simplify many tasks of machine learning. Thus, these examples are relatively simple compared with algorithms implmented with Python.