Traditional steps to implement ML
In this section, you will get a better understanding of the critical steps needed to produce an optimal ML model:
- Data preparation
- Machine learning model evaluation
Data preparation
A typical step in ML is to convert the raw data for input to train your model so that data scientists and data analysts can apply machine learning algorithms to the data. You may also hear the terms data wrangling or feature engineering.
This step is necessary since machine learning algorithms require inputs to be numbered. For example, you may need outliers or anomalies removed from your data. Also, you may need to fill in missing data values such as missing records for holidays. This helps to increase the accuracy of your model.
Additionally, it is important to ensure your training datasets are unbiased. Machine learning models learn from data and it is important that your training dataset has sufficient representation of demographic groups...
