Understanding feature engineering
ML is all about data. No matter how advanced our algorithm is, if the data is not correct or not enough, our model will not be able to perform as desired. Feature engineering transforms input data into features that are closely aligned with the model's objectives and converts data into a format that assists in model training.
Sometimes, there is data that may not be useful for a given training problem. How do we make sure that the algorithm is using only the right set of information? What about fields that are not individually useful, but when we apply a function to a group of fields, the data becomes particularly useful?
The act of making your data useful for the algorithm is called feature engineering. Most of the time, a data scientist's job is to find the right set of data for a given problem. Feature engineering requires knowledge of domain-specific techniques, and you will collaborate with business SMEs to better understand the...
