Classification models
Classification models sort new data into discrete buckets. You might use one to flag an incoming email as spam or not, or predict whether a customer will churn or stay. The model learns from past examples of each class in each case and then decides where new cases belong.
Clustering models
Clustering models look for natural groupings in your data without any labels. They’re great for customer segmentation—say, grouping shoppers by their buying habits—so you can tailor marketing or product offers to each segment’s unique preferences.
Regression models
Regression models forecast a continuous number, like sales revenue or temperature. For example, you could feed last year’s monthly sales into a regression model and get a prediction for next month’s sales figure. This helps with revenue recognition and inventory or staffing planning.
Time series models
Time series models focus on data that’s ordered by time. They pay special attention to trends, cycles, and seasonal patterns. For instance, you can predict next week’s website traffic by training on your daily visitor counts over the past year.
Decision trees
Decision trees split data step by step, asking a simple question at each node (e.g., “Is credit score above 700?”) until they arrive at a decision or value.
You can use them for both classification (like approving/denying a loan) and regression (predicting how much a loan will cost).
Neural networks
Neural networks use layers of interconnected neurons to learn extremely complex relationships in large datasets. They excel at tasks like advanced fraud detection or image-informed forecasting, where traditional models might miss subtle patterns.