Machine Learning 01

Introduction:

At first we need to know what Machine Learning(ML) is. Generally, people make a mistake by assuming that machine learning refers to learn from machine. Actually it's the opposite. ML means making the machine learn. Computers can make predictions and take decisions on their own with the help of ML. Imagine you are playing a coin toss game with your computer where you are tossing a two sided coin at each turn.  First turn is yours and it could be either head or tail. The next turn is the computer's and computer will not aware of the first turns outcome. After several turns if the coin is in head state then the computer wins and you win otherwise. So, there is 50% probability that you will win or the computer will win. Let's apply ML to this game. The first game will be played like before and there will be again 50% probability that you will win or the computer will win. Then the computer will record all of it's moves and label it's outcome with this information. After several games you will feel that computer is winning more times than before. Just a normal human player it is learning from it's experience. The more game it plays, the more information it gathers and the probability of it's winning increases. Imagine another game of rock-papers-scissors. The more a person plays with the computer, The more it gathers data and outcomes. Computer tries to find patterns from this data it gains and takes decisions accordingly. There are too many unstructured data around the world which is becoming very tough to store and process. Human writes thousands of lines of codes for every distinct situation. But using ML we can write code once and let the computer decide and manage data by itself.

Classification of Machine Learning :

Machine Learning can be classified into two major categories. They are :
1.Supervised Learning
2.Unsupervised Learning
There are two more categories namely Deep Learning and Reinforcement Learning. Those will be discussed in this blog later.