What is Machine Learning?
A machine learning algorithm is a computer program that is able to perform a task without being explicitly programmed to do so. Machine learning programs are trained to perform a certain task or make some prediction based on historical examples. The successful completion of the task or correct prediction will be measured as the performance metric. The learning program should be able to improve its performance metric with increase in historical example.
In supervised learning, we are given a dataset (training data) with expected result and we training the computer program using some algorithm based on the training data.
Supervised learning can be further classified into regression problem such as prediction housing prices, stock price and production volume etc. Supervised learning can also be classified into classification problem where the task of the computer program is to classified a problem into 2 or more categories. For example we can classified a tumour as malignant for benign.
In supervised learning, first we use predetermined dataset as training data. Then we choose and adopt a learning algorithm before we start training the learning algorithm with our data.
There are many machine learning algorithm. The most basic ML algorithm is Linear Regression. However, we do not apply linear regression very often since most things we want to predict are not in a linear form. Most commonly, we used polynomial regression algorithm.
We use Logistic Regression for classification problem. Besides Linear/Polynomial Regression and Logistic Regression, we also use Neural Networks and Support Vector Machine for supervised learning.
For unsupervised learning, we feed the computer program with raw data without expected or predetermined result. The program suppose to examine the raw data and define its structure by detecting similar features or characteristics.
For unsupervised learning, we use K-Means clustering and PCA (Principal Component Analysis). We also use Gaussian Abnormally Detection Algorithm to monitor and detects abnormality.
*** End ***