Stages in Machine Learning (ML)

Data improvement

Data collection and quality control.

Data from the real world or from opinion surveys.

Datasets that will be enriched as the analysis progresses.

This stage requires a lot of time.

Example:

Basic Statistical Descriptions

Properties of the data.

Pivot tables.

Business Intelligence

Identification of atypical values: outliers, fraud… or innovative profiles based on new topics to be soon incorporated as habitual behavior (business opportunities).

Prediction

Regression-type methodologies. Classifiers.

Linear & nonlinear approaches.

Hypothesis testing.

Machine learning.

No blackbox: no unknown formula.

Implementation

Algorithm in production.

On premises or in the cloud.

Involved stakeholders.

Outside intervention.