get-excelsior

[get-excelsior]

In this article, we are going to learn about Linear Regression 

Let’s first understand what Machine Learning is. It is a technique by which systems like computers are able to learn and adapt without any explicit knowledge given, by utilising the algorithms and some models to study and get information from the data. In a broader sense, it is a branch of Artificial Intelligence. Highly used product, Alexa is based on data and Machine learning which supports its usage in today’s world. 

Linear Regression is a supervised machine learning algorithm that works on the principle of straight line. As by its name, it is a regression technique where output has no labels defined. The goal here is to predict the continuous range/values. Example: stock price, house price etc. Here we can’t predict discrete values, hence regression is used. 

Similarly, in Linear regression a continuous spectrum of output is predicted by the model that has values nearer to the actual output. In this algorithm, a small amount of the big data is given to train it and learn the attributes so that when the rest of the data is given to it, outputs can be predicted accordingly. 

Linear regression assumes that there is a linear relationship between the input data(x) and output data(y). 

Features are the independent variables of the dataset that can be represented by x1, x2, x3, x4… xn, where n is the number of total features. 

Hence, the linear relationship between the x and y can be shown as :

Y = m1x1 + m2x2 + m3x3 + … + mnxn + b 

Where Y is the output, m are the parameters and b is the intercept of line

When we pass the training data to the algorithm it somehow tries to learn these parameters and the value of b to find the best fit line to get the minimum errors and predict the output of testing data.  The smaller the error, the greater is the accuracy of the regression model. 

visit us: www.getexcelsior.com