Ad Code

Machine learning in simple words

☆What is Machine Learning 

Machine learning is a subset of Artificial Intelligence which provides machines the ability to learn automatically and improve with experience without being explicitly programmed to do so.

But there is question??
How can a machine think or make decision?
These machine are very similar to the humans. If you feed a good data to machine. It interpret, process and analyze the data by using algorithms.

☆ Terms using in Machine Learning 

▪︎ Algorithms 

Algorithms are the set of rules that is used to learn patterns from data and provide significant information. 
Linear regression, decision tree, random forest define the logic behind machine learning model.

▪︎ Model 

Model is the main component of machine learning process. It is use in order to draw useful insights from the input and give you a outcome. 

▪︎ Predictor variable 

The next we have is predictor variable. It is used to predict the output. It is the part of data that we need to predict the outcome. 

▪︎ Response variable 

The another is response variable. It is also known as target variable or output variable. This is the variable that you want to predict by using the predictor variables. 

▪︎ Training data 

The another term is training data. Training data is used to create model in machine learning. When you feed the data in machine during the machine learning process. The data divided into two parts that is training data or testing data. It helps to the machine to identify the patterns. 

▪︎ Testing data

The another term is testing data. We create machine learning model. We have to divide the data into two parts that is training data or testing data. So testing data is used in model after machine is trained. We test the accuracy of the outcome here after the training of the model. So it is the use of testing data. 

So I have discussed the terms of the machine learning. Now let's see the process of machine learning.

☆ process of Machine Learning

1. Define objective 

The first step in machine learning process is define a objective.
I give you an example so that you can understand it better. Let assume that you want to predict the weather condition of our locality. So firstly, you need to answer questions such as what are we trying to predict. Our target variables will be tell us that whether it's going to rain or not. Input data is we will need to predict such as temperature of the location or the humidity level and so on. 
So you need to define the objective. 
We need to predict target variables and what are the different predictor variables. So basically you have to form an idea of the problem. 

Let's move on the second step of process.

2. Data gathering 

In this stage, we need to know questions such as what kind of data is needed to solve this problem? And is this data available? And if it available from where can i get this? 
Data gathering is one of the most time consuming process in steps of machine learning process. If you collect the data manually it will take lot of time. But we are lucky because we have a lot of resources online which provide data set. 
Now we will jump to our example. For predicting the weather.  we need data for weather forecasting. We have to humidity level of the location. We have to collect such kind of data or stored for analysis. So all the data is collected during this stage.

Let's move on the third step of process.

3. Preparing data 

The step is followed by data preparation. It is also known as data cleaning. If you see the data which is collected by you it is not almost in right format. If you are taking data from any website, the data will require more cleaning and preparation. You have to filter the data so that it will ready for analysis. 
You need to remember some things during cleaning of data like you should find missing values, redundant variables, duplicate values etc. 
This was the most important step you need to remember it. Most of the data scientist said that it is the most important and time consuming step of machine learning process. They also said that 80% of their time is consumed by data cleaning. 
So it is not easy to get rid of missing values or corrupted data. Your data might get affected during this step. 

Let's move on the fourth step of process.

4. Data exploration 

The next step is data exploration or exploratory data analysis. In this stage we are a detective. It is a brainstorming stage of machine learning. 
Data exploration involves understanding of patterns and tends in your data. In this stage, all the useful insights are drawn.
What is the meaning of patterns here? 
Let me explain, let's go back on our example which we have to predict the rainfall. As we know it will depend on the temperature or the humidity level of location. The outcome will be depend on the variables. We have to find out patterns, correlations between such variables. 
It is the most important part of machine learning. 
You understand what exactly your data is and how you can form the solution of the problem. 

Let's move on the fifth step of machine learning process.

5. Building a model 

The next step in the machine learning process is Building a model. All the patterns and insights that you derive the data exploration are used to build a machine learning model. There are slipt in data set in two parts which are training data and testing data. 
I discussed training data and testing data in Terms of Machine Learning.
When you building a model you always use the training data. Now you will think what is training data? So training data is the same input data that you are feeding in the machine. The only is you used 80% of you data for training purpose and 20% of data for testing purpose. 
Basically model is used the machine learning algorithms that predicts the output. The algorithms are bassed on your problem that you are trying to solve. Because there are n numbers of algorithms so you have to choose algorithms according to your need or problem statement. 

Let's move on the sixth step of machine learning process.

6. Model evaluation 

The sixth step of machine learning process is Model evaluation. After you have done building a model by using the training data set. Now it is time to check the model. The testing data set is used to check the efficiency of the model and how accurately it can predict the outcome. Once accuracy is calculated any further improvement can be implemented during this stage. The various methods can help you to improve the performance of the model. Like you can use parameter tuning and cross validation methods in order to improve the performance of the model. 

The main thing that you should remember during model evaluation and optimization is that model evaluation is nothing but you are testing how well your model can predict the outcome. So in this stage, we use testing data set. In the previous step we use training to data set to build a model. In this step we use testing data set. You need to calculate the accuracy of the model.

Let's move on to the next step of machine learning process.

7. Predictions 

This is the final step of the machine learning process. Now, once a model is evaluated and once you have improved it. It is finally used to make predictctions. The final output can either be a categorical variable or a continuous variable. All of this is depends on your problem statement. Let's go on the example if we are predicting the rainfall. It will be categorical variable. 

So that was the entire machine learning process. Now let's move on the types of machine learning. 

☆ Types of Machine Learning 

1. Supervised Learning 

It is basically a technique in which we train the machine by using the data which is labeled. In order to understand supervised Learning let's consider a small example. So we all need to solve mathematical problems. A lot of us trouble solving these problems. So our teachers helps us to understand the problem. Similarly you can think about supervised learning that involves a guide. The label data set is a teacher here that trains to understand the patterns of data. You need to label data that you feeding to the computer. basically your entire training data should be labeled. That how supervised learning works. 
One thing that you should note down here is that the output you received after the process of machine learning in supervised learning is also labeled. 

Let's move on the next type of machine learning

2. Unsupervised Learning 

Unsupervised Learning involves training by using unlabeled data. I have discussed with you in supervised learning we use labeled data. But here we use unlabeled data to train the machine. Here we allow the model to act on the information without any guidance. There is no supervision here. Basically it means learning without any guidance. 
In this type of machine learning, the model is not feed in any labeled data. The model does it will from different clusters. 
So the important thing that you need to know in unsupervised learning is that you feed the machine unlabeled data. 

Now let's move on the last type of machine learning. 

3. Reinforcement Learning 

The last type of machine learning is Reinforcement Learning. It is different from the supervised and unsupervised learning. In reinforcement learning, machine learn how to behave in the environment by performing certain actions and observing all the things. 
Let's take an example, imagine that you lost in an island. So you need to know about the environment of the island and observing everything to survive there. You have to find what is growing there to eat and all about the dangers there. 
Basically reinforcement learning means identity the environment and start observing things that result in rewards.

So, I have discussed all the types of machine learning with you that is Supervised Learning,  Unsupervised Learning and Reinforcement Learning.

Here is my attempt to explain Machine learning in simple words. I hope you understand it well. 

Post a Comment