What is Machine Learning?
Machine learning uses a set to tools and algorithms to make connections between individuals and their behaviours. These routines can be TRAINED to look for predictive clusters of behaviour. So here is a brief introduction to machine learning ideas.
We can infer by analysis of the data that individual A – let’s call him Alfred – has sufficient correlating indicators in his behaviour that he is likely to carry on buying from us. Individual B – let’s call her Barbara – displays correlations which suggest she is more likely to churn, to move to another company’s offers.
Minimizing churn is a key way of maximising the long term shareholder value of an organisation. This suggests two possible routes forward for action
- Alter our marketing narratives and activities to attract more Alfreds.
- Identify offers that overcome Barbara’s tendency to churn
Or of course both.
Churn is one of the areas that we will be looking at in this course and we will explore how the process of CLASSIFICATION can help us here.
So what is machine learning?
This brief introduction to machine learning covers the following points
- Customer segmentation clusters your customers so you can identify groups and see how similar customers behave
- Churn prediction compares your subscribers/customers to previous ones, and detects who is about to leave by classifying them.
- Prediction is often carried out by using Regression techiques which use mathematical floating point formulas instead of Boolean operators (AND, OR etc). For example I once created a regression model to predict steel demand in the UK economy. It was based on historical data of the 4 biggest consuming sectors (construction, motors etc) and import and export data. If you entered assumptions for the independent variables it would tell you what the demand for steel would be.
- Rule extraction. The search for patterns in data. A good topical example of this is the success of the Brexit and Trump campaigns which used Social Media to identify potential supporters. They were then selectively advertised to with material that appealed to their core motivations. During the course we shall explore how businesses can take advantage of such algorithmic use of big data to expand their advertising target market.
What we will cover in this course
We will be looking in detail at decision trees, cluster analysis and support vector machines and these are reviewed in more detail here.
We have tried to integrate data science and machine learning into the course as a whole as this diagram illustrates. This should provide a suitable introduction to machine learning
To find out more
Learn more about Machine Learning Applications
Consider downloading our Digital Marketing Papers