At a high level, machine learning is a way for us to answer questions that are difficult for us the answer with human intellect. We use computers instead that can answer questions and do things much faster. They can do things interactively and over large data sets and come up with answers that would take us forever to come up with.
What type of questions can machine learning answer for us? Let’s look at 5 basic questions:
1. How much or how many? So, for forecasting or also called a regression test. How much money am I going to make next month in which district for one particular product?
2.Which category? What category does ‘x’ fall into? A common use case would be for sentiment analysis, so to determine which category a tweet falls into within sentiment. If a tweet is mentioning your company, is it good or bad?
3. What group does this fall into? This would be some sort of clustering algorithm. Maybe we’re looking at our customers or medical patients as a whole. How do they group together? So, looking at all their attributes and how they cluster together and fall into groups. This will allow us to think about how to approach those customers or patients going forward.
4. Is this weird or is something not normal? An anomaly detection. A perfect scenario is if you’re using your credit card in a state other than where you reside, and your credit card company calls and says, this is weird, should your card still be working?
5. What options should we take? This would be a recommendation model and based off the history of the data we have, what is the best option going forward? A common scenario would be a product recommendation on a website such as Amazon.
So, the first step of any machine learning experiment is deciding what questions you want answered. As you dig deeper into machine learning, you’ll be able to see which model to use.
If you have any questions on how to jump start your machine learning project or how to get some data science initiatives going with your business, we’re here to help. Click the link below or contact us to get a discussion going—we are your best resource.