How Machine Learning is Transforming Data Strategies

How Machine Learning is Transforming Data Strategies

Before we delve deeper into machine learning let’s talk about what it is. I will give you my definition of machine learning since, believe it or not, I have heard machine learning defined in so many different ways by so many different people.

I define machine learning as the methodology of programming computers to act like humans do, and improve their learning autonomous over time, by feeding the programs data and information in the form of observations and real world interactions.


Now, let’s give you a firm understanding of what the basic concepts of machine learning are and how they can help your business intelligence solution.

There are many, many different machine learning algorithms and hundreds being published every day. While machine learning has been around for while, we’ve only just started to see that some of the biggest known examples of machine learning have been announced. The Google self-driving car, knowing what your customers are saying about you in social media, and fraud detection which is being used all over the world. These are just some recent examples. There are many more where machine learning is applied to help consumers; I’m going to briefly walk through two algorithms that can make a big impact.

1. Classification and Regression Trees or CART (Decision Tree): This algorithm classifies data into different groups using certain attributes, the data is pushed thru the model and at every decision point a decision will be made in what group it falls into.  Let’s say we have data coming into the algorithm, how many wheels a vehicle has, and how heavy is the vehicle. We then can categorize or classify the data into groups based on the attributes.

Wheel = 2 then motorcycle

Wheel = 4 and weight < 4000 then car

Wheel = 4 and weight > 4000 then truck.

2. Naive Bayes: This algorithm extracts pieces of text from other text and examines the attitude is it positive or negative. Let’s give some text as an example and look on how it gets processed and received a sentiment score.


Hi Bob,

It was nice speaking with you yesterday. Here is the link to the updated individual dash:

I added the new component, Closed Business LTM by Account, and I updated the other components to make sure the date filters were correct. Please let me know if this dashboard looks good and then I can get started cloning them for the other users.

Thank you!


Once we ran the above text thru the algorithm in Amazon Web Services (AWS) we received back a score: “Sentiment”: “POSITIVE”, “SentimentScore”: {“Mixed”: 0.010181277990341187, “Negative”: 0.00235496973618865, “Neutral”: 0.06723286211490631, “Positive”: 0.9202309250831604}}. This score represents the likelihood that the sentiment was correctly detected. For example, in our example above it is ninety-two percent likely that the text has a Positive sentiment. There is a less than 1 percent likelihood that the text has a Negative sentiment. You can use the sentiment Score to determine if the accuracy of the detection meets the needs of your application.

These are just two of so many machine learning algorithms that you and your business can leverage to improve your business intelligence solution. Having the right team in place – whether it is an internal team or an external consulting team – is important because experience and knowledge of machine learning techniques are critical to the success of implementing machine learning into your business.

Contact us if you’d like to learn more about machine learning and business intelligence.