Feature Engineering

Machine Learning algorithms use input data to create output/model.This input data comprise features(attributes/column names), which are usually in the form of structured column.

Feature Engineering is fundamental to the application of machine learning.

Feature Engineering

Feature Engineering is the key to improve the performance of Machine Learning models. Feature Engineering is the process of transforming raw data in the form that represent the problem in hand. Feature engineering are covered beginning with data wrangling which will help to explore techniques for collecting data over the internet. Once the data is collected most of the features may not be relevant or the features may be correlated to overcome such issues before modelling we cover dimension reduction and feature selection. In some cases, the features collected may be in different ranges hence we cover Preprocessing techniques to normalize the features followed by visualization.

Feature Engineering creates new input feature from existing. This includes data cleanup, aggregartion(merging multiple column) and picking data randamly from collections.

Algorithms requires features with some specific characteristic to work properly.

two goals
  • Preparing the proper input dataset compatible with the machine learning algorithm requirement.
  • Improving the performance of machine learning models.

Output/model decides based on data, that's why datascientist spends 80% tome to validate/prepare that.