NEW YORK – MicroStrategy’s out-of-the-box machine learning (ML) algorithms, extensive built-in function library and native R integration can all be leveraged by organizations to expand the depth and scope of their analysis, according to Vasant Paranjpe, MicroStrategy senior sales engineer.
“The one thing I’ve learned when trying to keep my head around this whole machine learning space is it changes every day,” he said Oct. 25, during a conference session called “Machine Leaning at MicroStrategy: Injecting sophisticated data science into real-world applications,” at the MicroStrategy Symposium, held at Convene’s downtown Manhattan event venue in the Financial District.
There are “three classes” of ML out there: supervised, unsupervised and reinforcement, he pointed out.
Paranjpe explained: “The unsupervised side is when the machine learning models just kind of go out there and do their own things and keep learning on their own. The supervised one is what we’re really talking about in the demo that we’re doing here [and is when] I have a known output that I’m trying to predict. I want to make sure that I give the algorithm the right information and the algorithm is able to predict those [specific] things out there. And, from what I’m told, the reinforcement side” includes all the “cutting edge” and “next-generation things on process.”
The majority of ML problems are supervised and require labeled data in which the outcome is known, according to MicroStrategy. While unsupervised ML is used to discover patterns that exist in data, one problem is that there’s no way to test accuracy, the company said. Reinforcement ML, meanwhile, is used today in game playing such as the AlphaGo computer program developed by the Alphabet-owned AI company DeepMind Technologies that plays the board game Go and in manufacturing including Amazon warehouse robots, according to MicroStrategy.
Advantages of leveraging MicroStrategy in an organization’s ML workflows include: the fact that it’s a single comprehensive platform to centralize business data used for ML; it provides one version of the truth that helps protect against data scientists training models on the wrong data; it offers enterprise grade self-service that empowers business users and data scientists to work together; and it enables an organization’s workforce to take action on predictive workflows wherever they are, according to Paranjpe.
He told attendees: “If you’ve been to a lot of MicroStrategy Symposiums before…enabling the workforce has really been a key theme of how our customers use MicroStrategy… Whether it’s putting a mobile app in the hands of a salesperson, whether it’s putting an executive dashboard in front of an executive, if it’s putting the right information in front of a buyer or purchase department. It’s really making sure that workforce can use all this data that’s through the enterprise in a real way and a tangible way.”
Another major theme at the Symposium was how artificial intelligence (AI) can play an important role in organizations successfully transforming into intelligent enterprises, although it can’t be relied on to cure all of a company’s problems.