What Is Machine Learning
Posted By : Rajat khurana | 25-Jan-2018
Instead of Learning machine learning, we will know what machine learning does internally?
1) Find Patterns in Data.
It finds different patterns in data by analysis of different algorithm and based on some other criteria.
2) By using those patterns to predict the future. For eg, we have a lot of data related to the different organization then we will find patterns and arrange them accordingly like service or product based company, then there is a different subcategory in it.
Thus machine learning is totally based on data i.e recognizing the pattern, for it, you will need to become a Data Scientist to analyze all data. A simple example of machine learning is credit card fraud which will be traced only by analyze data and generate a pattern from it.
So, for machine learning you need Data that contain patterns, pass these data into machine learning algorithm and analyze the pattern. This process will generate a model which recognize a pattern. Thus the most important thing in Machine Learning is "Model". The different application uses these model and passes different data to recognize the pattern. Machine learning process is very easy
Now come to points, how we create an application in machine learning. There are a lot of languages available for machine learning but the extreme language is "R" and "python". R is an open source programming language which supports machine learning language and more. It has many available packages which are used in machine learning and same is for python.
The most important thing in machine learning is that you need to update Model after some time, this time vary depending on the application. This process is iterative.
There are different terminology used in Machine Learning
1) Traning Data: Prepare a data to create a model which is used in applications.So, creating a model is training a model.
2)Supervised Learning: In this value that you want to predict is in training data. So data is labeled.This is the most commonly used in machine learning.
3) Unsupervised Learning: In this value that you predict is not in training data. So, data is unlabeled.
There is a different category of machine learning problems like Regression, Classification, and Clustering. We will discuss it later on and create an application in Machine learning.