# But how to start machine learning?

But how to start machine learning? In machine learning is a widely popular phrase in the current computer world. So, many researchers, developer, and a mathematician are involved in research on AI(Artificial Intelligence). So many new computer science students want to involve in machine learning.

There are different applications of machine learning. What is Machine Learning is already explained here.

But, there is confusion about how to exactly start machine learning?

First this you need to know that machine learning is all about computation theory, mathematics, and optimization, so we need to learn different theories and math. Artificial intelligence development (aidev) is tricky and something smart at all. Here I explain some important tips to start machine learning.

##### Statistics

The huge amount of data is used to make machine intelligence. We need to proceed with large scale data. We need a data-based decision. Statistics is a math of data. It is able to calculate big data to make a certain decision. There are different approaches to statistics to make machine learning possible.

1. Linear Regression
2. Logistic Regression
3. Correlation
4. Kurtosis Analysis of Data
##### Probability

Machine Learning is based on probability theory. It’s all about may or might i.e Some event may/might occur. Machine Learning algorithms make some assumptions based on probability theory. Naive Bayes is one of the best example which is work based on probability theory. So you need to know about probability.

##### Linear Algebra

The computer is not able to understand language (words, phrase) we need to change into a numeric value which gives a big series of numbers in the matrix form. Linear algebra used to manipulate such matrix and calculate different matrix calculations. For matrix operation we need to know linear algebra.

##### Calculus

Calculus is a math of optimization. An algorithm must be fast and optimum. Generally (derivative) is used to analyze algorithms and it’s the best optimum way. Machine learning is used calculus of optimization. Different optimization algorithms like gradient descent, hessian are an application of derivative in machine learning.

##### Algorithm

There are different algorithms in machine learning. All have their properties and won working conditions. Algorithm analysis in machine learning is used to choose the best algorithms for our working module.

Some of the algorithm listed below:

1. Naive Bayes
2. Decision Tree
3. Dimensionality Reduction
4. Linear Regression
5. Logistic Regression
6. Support Vector Machine
7. Backpropagation

Above mentioned algorithms are used according to our requirements.

##### Programming Language

To make certain applications or change our theoretical requirement into a model we need to do some code. There is no restriction of the programming language we use either c, c++ or python. But I personally recommend python language because of the highly supportive towards machine learning and rich libraries.