We make choices everyday and our decisions are based on certain factors. If we can keep records of what factors we consider and their respective values that we use to make a decision, we observe certain trends which we may be able to use predict what choice one will make given some values of the factors considered.
This is a typical supervised learning task which can be achieved computationally by the probabilistic algorithm known as the Naive Bayes algorithm.
In Ghana, political inclination may depend on ethnicity, level of education, current economic state, occupation, gender, etc. Thus if we wanted to tell one's political inclination given that he comes from the Volta region, has basic education, fair economic state, carpenter and male, we have to consider all these factors and this changes the dimension of the Bayes' formula.
Let us work with this example from a U.S setting. The goal of the project is to predict the political inclination (conservative or liberal) of a person based on his or her job (analyst, barista, cook, or doctor), sex (male, female), and annual income (low, medium, high).
Sample input data
Classification
Accuracy would be significantly improved with more data