Functional Groups

Functional Groups are those organizations and clusters of professionals that participate in Machine Learning. There are a number of ways to organize functional groups involved in Machine Learning.

Functional Groups in this context is not meant to be applied exclusively to formal reporting structures. It can also be viewed as collections of expertise.

A Framework for Machine Learning Functionality

The graph diagram below portrays an approach to organizing ML functionality. You can view the diagram as a tool for thinking through your own individual plans and strategies.

The function shown in the five cluster circles (Mining, Programming, Analysis, Reporting, Modeling) are blended in different ways by different organizations. For example, the Data Cleaning function included in Mining below might be included in Analysis in a differently conceived structure.

In fact, and in practice, the five clusters support each other and interact heavily in the overall Machine Learning process. This is why they are shown as interconnected to one another.

Mining

Other terms applied to this category include: data pipeline, data wrangling, business intelligence.

Expertise in this area includes knowledge of:

  • data and database structures

  • data query languages

  • data sources

  • business intelligence

Programming

Other terms applied to this category include: coding, system development.

Expertise in this area includes knowledge of:

Analysis

Other terms applied to this category include: data analysis, information processing.

Expertise in this area includes knowledge of:

Reporting

Other terms applied to this category include: data reporting, data collection, information display.

Expertise in this area includes knowledge of:

  • business requirements

  • reporting formats

  • visualization options

  • communication effectiveness

Modeling

Other terms applied to this category include: data pipeline, data wrangling, database management.

Expertise in this area includes knowledge of:

Additional Factors to Consider

Connection to Client Organizations

Machine Learning often serves the purposes of a variety of client organization. One way to maintain an effective connection between a centralized Machine Learning organization and client organizations is to embed select ML experts in client organizations. This can leverage the efficiency of a central organization while maintaining responsiveness to client organization needs.

Flexibility

The needs of individual Machine Learning projects can vary.

Agile Processes

Agile processes include cross functional members. In the context of the diagram above, an agile team might include experts from many functional areas, such as mining, programming, analysis, reporting, and modeling,.