Ensemble models
We will build multiple algorithms and combine the results of different algorithms with different weightages. We can decide on the weightage based on a trial-and-error basis. As discussed previously, the dataset can be divided into a training set and testing set, and we can evaluate the performance with different weightages.
The other popular regression models that can be implemented are Support Vector Machine (SVM) and Random Forest. The SVM algorithm is where the model is built by constructing a hyperplane, and in random forest, we build the model by building a number of decision trees.
Replacing NA with mean or median
When there are very few records with blank or NA values, then we can also consider replacing them with the average value. This methodology may or may not help in improving the accuracy, and so, it is important to test the performance on a sample test data. This can be implemented using the following code.
It is preferred to replace the missing values with the...