Linear Regression vs Classification

Linear Regression:

Well, the term sounds a bit mathematical for a person encircled with non-Machine Learning stuff. Don’t go on it’s name either, as the algorithm is as easy as walking in a park. It follows a simple paradigm- “prediction on the basis of previous trends“. Alternatively, it’s something that tends to go with a flow. Suppose, you have two variables, one on the x-axis and the other undoubtedly on the y-axis. You have plotted like thousands of points on the plane for co-ordinates. The plotted points will of course be following some trend. In simpler words, the points must be following some kind of pattern. The regression algorithm tends to draw a line that goes along with the trend and predicts the values of dependant variables (on y-axis) from the points that we input as independant variables(on x-axis). It basically tells the y co-ordinate for some x co-ordinate that are present on that particular line. One more thing, the algorithm follows a continuous trend, means the output can be infinite variety of numbers.

Classification:

Now comes another crucial algorithm in the field of Machine Learning. The name is quite self-explanatory though. It classifies the input into any of the pre-defined classes. For instance, you have two images, one of a cat and a dog. Suppose you provide an image of a dog and ask the machine what class does it belong to. It will figure out the features of the input image and compare it with both pre-defined classes and ultimately will give the output with the class the machine has predicted.

Well ! These are just quite basic concepts of regression and classification. There comes more sub-categories for these algorithms, how they work in some different scenarios and much more.