Regression_Model

Machines are human until we train them like human!

Regression: It is the part of Supervised Machine Learning model used to feature the raw data into real output variable(Y). Where we have input variables(X) and output variable. Here we use algorithm to map function from input to output Y= f(X) . 

 Regression model output predicts real or continuous values. This model simply tries to fit data from real hyper-plane mean normal X- Y quadrant plane by going through different points in the quadrant which helps to predict real dependent variable from wide range of in dependent variables.

This model is so power that we can predict price of any house in the particular city based on available raw data. Based on different attributes we can find out anyone's age , salary , nationality , stock prices and all.

Toughest part of regression model is feature engineering this is simply the iteration to determine anticipated data to train model and converting raw data from different range like log files, CSV files into real predicted output. As to modeling this we can have different frameworks like Tensor flow , PyCharm and many.

Feature: A feature is an input variable , the x- variable in simple linear regression . This is just like particular instance of data. we can use millions of feature.

 {x1,x2,x3,...........xn}.

Labels: A label is predicting the Y- variable in simple linear regression . The label could be price, age , nationality.

Labeled examples: It includes both features and labels.

                { features, label } : (x,y).

Unlabeled examples: It contains features and not the labels.

                  { features : ? } : (x,y).

Supervised Learning is best described in and as in Linear regression. This is just finding the straight line or hyper plane that best fits a set of points in the x-y coordinate.

Let's consider a linear equation where we try to predict the output based on inputs.

y = mx + b:

• y-> temperature trying to predict in Celsius.
• m-> slope of the line.
• x-> the number of temperature variables.
• b-> it is y intercept.

In this case if the line does not passes through every dot line and based of distance from line to points we can determine losses and normalize it which closely gives relationship between output and input in terms of function.

This is bit of Regression model which i tried to put it on as much simple that i can do.

Hope you enjoyed my artifact.

Until then await for my next repertoire!