Data Science has been at the core of our parent company Macrosoft Inc. throughout its 25-year history. This specialization became the foundation for the creation and success of Macrosoft AI. What makes Macrosoft AI special is that our proficiency and focus on Data Science is on par with that of our AI model development/evaluation and deployment capability. We have been adherents of the CRISP-DM methodology (a cross-industry approach to data mining) for the past decade, on both AI engagements as well as with other technologies such as database development, business intelligence and Big Data solutions. This methodology is an idealized sequence of events as depicted in the image below. Over time, we have learned that two of these events – Data Understanding & Data Preparation – require a degree of customization when using this methodology for AI.
Data Preparation Methodology
Data conclusions drawn at the understanding step is led by gap analysis and creating derived attributes. This in turn leads to AI-powered Gap Remediation where completely new data can be generated from the data-on-hand.
Data Understanding Methodology
One of our guiding principals at Macrosoft AI is to become expert on our client’s internal data as well as any complementary sources that are determined to be relevant for the use cases. We apply machine learning algorithms for data profiling and classification, as well to draw conclusions from the data.