7.10: Conclusion

We discussed different feature selection methods in this chapter. These methods can help us improve the model performance marginally. We cannot expect these methods to provide the best possible models. This is especially true when the model performance is already low. Most of the methods discussed in this chapter are not computationally expensive. The ideal scenario for using these techniques will be when we already have a well performing model and the number of features are very high. We can obtain a marginally better performing model with a smaller number of features with these methods.