Feature Engineering Within the Predictive Analytics Process — Part Two

In the last article, I discussed the concept of feature engineering as comprising two components with the first component being the ability to create and derive meaningful variables in the analytical file which is used as the source information in the development of any predictive analytics solution. Within this first component, access to data has grown exponentially with data scientists now being able to use semi-structured and unstructured data as data inputs. Although the ETL process for this type of data requires new technical skills to essentially structure this data, the intensive data science work in creating and deriving variables still adheres to the same disciplines that have been used for the last 30 years. Filtering/Reducing the Variable Set using Factor Analysis Meanwhile, the second component looks at tools and…


Link to Full Article: Feature Engineering Within the Predictive Analytics Process — Part Two