As we have seen by now that the data set can be non-linear. I mean to say that if I apply linear regression then we may not be able to fit the line very well.
For the case of data which had 1 variable, it was possible to visualize the data by plotting it on the graph. But in real scenarios, we may get data with many features. In that case, how do we decide which algorithm is to be applied?
Linear Regression
Locally Weighted Regression
Logistic Regression
.
.
other Future algorithms in the course?
Which model to choose?
Hello Sagnik,
That’s a really good questions.
Actually, the modles we use depends completely on the kind of data we have. For example, the two broad classifications of ML models as supervised and unsupervised is also differentiated by the kind of data we have. Supervised models are used when we have labelled data.
Similarly, depending upon the kind of data set, number of features, data size and number of variables we choose a model.
As you progress through the course, you’ll see a lot of ML algorithms which are used for specific purposes and are most efficient for some kind of data. So stay tuned, you will get an answer to this question automatically in the coming days as you progress through the course.
I hope this helps.
1 Like
Well thanks for answering my question spontaneously. Now I am more excited to progress in this course. 
1 Like
That’s great, pleasure to help. Glad to hear that 
1 Like