2 February, 2017
Machine Learning And Analytics: What’s Your First Step?
Machine learning is a growing field, used in everything from the basics of anti-spam functions to the complexities of self-driving cars. As this is a constantly adapting technology, companies seeking to take advantage of the system for functions like analytics may have trouble finding the best place to begin.
So what is the first step for a tech department that wants to start using machine learning to improve its data analytics? Forbes Technology Council members have this to say:
1. Learn The Science
Even if you rely on outside expertise, it is important to understand what machine learning can and can’t do with data. Stanford, Caltech and others offer online classes on Coursera that are very good. Even a little knowledge will go a long way towards helping you identify consultants and opportunities for analysis. – Manuel Vellon, Level 11
2. Start With Basic Connectivity And Data Collection
The road to advanced analytics and machine learning starts with basic connectivity and data collection. This journey includes pinpointing the questions that need to be answered with data analysis, identifying the data needed to answer those questions, and putting processes in place to gather the correct type and amount of that data to properly support machine learning. – Mark Benson, Exosite, LLC
3. Have A Clear Goal
Ensure that you have a goal for the analytics. You need to train the machine. Ensure that you have the data and understanding of what the machine should learn, or it will be difficult to be successful. However, that doesn’t mean you have to have a fully detailed plan either. The best project outcome may be that there’s a positive surprise in the analytics. – George McKevitt, Compliance Science
4. Start With An Internal Project
Find an interesting problem that the team wants to crack, and let them develop their skills in machine learning while they work on a problem they are passionate about. – Kurt Dykema, Twisthink
5. Hire An Expert
Get an expert data scientist to work with the tech department and business sponsor. Identify a small, “low-hanging fruit” first project that can show true business value. Choose a project where there is a sufficient amount of data and a well-defined outcome that can be demonstrated. (The data scientist can then validate the data’s applicability.) – Leon Hounshell, Greenwave Systems