# 7 simple steps for mastering Machine Learning — for anyone

This post was originally published by Venkata Sai Krishna M at Medium [AI]

### What is machine learning in Simple Terms?

Before going there, let’s answer this question, if one of your colleagues, Krishna, is late for his work daily by 30 min, and you know it.

After some days of watching him come late for work by 30 min, a third colleague, Venkat, came to you to ask you for Krishna, what will you answer?

“Krishna will be late by 30 min!” right?

My point here is, based on your past experience or data you were trying to predict the future of Krishna’s arrival. Because everything has a trend!

Now assume if the computer does the same thing!! Based on the previous data if a computer is trying to predict the future, this is a classic example of Machine Learning.

#### And a million-dollar question is where to start?

According to Dale’s “Cone of Experience,” simulating and doing the real thing will give more learning than reading and hearing to lectures. All you need is a chance to simulate what you have learned.

To do that you need to work on this 1st step:

Step 1: Learn a programming language

Based on a survey in 2019 on engineering students, 58% showed interest in learning Python, 42% showed interest in R Programming, and 25% in both the programming languages.

For the purpose of machine learning and Artificial Intelligence, I would personally recommend you to go with Python or R Programming based on your comfort.

If you don’t know both, go with Python!!! (As I said it’s my personal recommendation no commercial benefits)

Step 2: Data Visualization

Usually, data will be represented in the form of tables, data frames, etc., but these are descriptive ways of representation. It’s hard to make anything deductions out of it (Unless you are a computer 😊).

To gain the capability of deductions out of the data you will have to learn Data Visualization to structure the data into graphs.

Step 3: Data Preprocessing

Data you receive in real-time may not always be clean or organized, problems like missing data and error values are sometimes inevitable.

Handling categorical variables like Gender, Names of Months, List of states, etc., So that system won’t think, one is higher and the other is lower.

Splitting the Data set into a training set and test set. Training set to make understanding of the data and Testing set to make predictions and check correctness before finalizing the model.

Last but not the least of the part is feature scaling, a mechanism to limits the values in variables to the comparative range across the data set, so that, variables can be compared with each other.

Till here is just a prerequisite for Machine Learning.

The Next step will kick start your machine learning

Step 4: Regression

With the help of past data, create a trend line for predicting the future

Simple Linear Regression will predict future values with 1 variable

– E.g.: Bus Timings of Krishna from our earlier example

Multiple Linear Regression will predict future values with multiple variables

– E.g.: Bus Timings, Traffic in the way of Krishna from our earlier example

Polynomial regression will predict future values with dependent variables

– E.g.: Bus Timings as well as the crowd in the bus of Krishna

Step 5: Classification

With the help of variable past data, if you are able to categorize, then its classification. (E.g., Only on Sundays “Bus Number 123” is crowded. Predict could be if it is a Sunday, “Bus Number 123” will be crowded.)

K Nearest Neighbors, creates a boundary for each similar group to predict the future values

Support Vector Machine, creates boundary on the trend line to fit between the groups

Decision Tree, creates a set of questions with multiple choice answers and grouping the values with common answers

Random Forest Classification, creates group of decision trees and create groups with common answers

Step 6: Association Rule Learning

This method helps to discover or map interesting relationships between variables in the large datasets

Apriori Method, identifies the frequent and common items in the large datasets

E.g. Predicting the purchase of Jam on the purchase of Bread

Eclat Method, also works like an Apriori Method but for small or medium datasets

Step 7: Natural Language Processing (The most interesting part of Machine Learning)

Make computers understand the human language by reading, deciphering, and making sense of the words just like a human.

Best examples are Google Assistant, Alexa, Siri, etc.,

Spam Filters, Spell check and auto complete texts are some of the examples

Conclusion

Once you reach this step nothing can stop you from landing in a job as Machine Learning Expert.

Thanks for ByteFrame Technologies for letting me download all the resources required for this topic.

Your next steps to becoming an expert with include Artificial Intelligence domains like Deep Learning and OpenCV concepts

HAPPY LEARNING!!! Happy to help you more!!!