What Machine Learning problems does the Engine tackle?

This article briefly describes the Machine Learning problems that the AI & Analytic Engine tackles.

The AI & Analytics Engine currently supports three types of Machine Learning problems, namely classification, regression, and time-series forecasting.

Classification (Supervised Learning)

Given an observation, predict which category something belongs to, from a set of categories.

Examples:

  • Binary classification - If an email is spam or not? (yes/no)
  • Multi-class classification - Which flower species does the observation belong to? (Iris setosa, Iris virginica, or Iris versicolor)

Regression (Supervised Learning)

Given an observation, predict the continuous variable.

Examples:

  • Predict house prices or taxi trip duration

Tip: Still curious? For a detailed comparison between classification and regression, see this article.

Time-Series Forecasting

Forecast numeric values such as customer demands and stock prices in the future, based on historical time-stamped data.

Examples:

  • Forecast stock prices for the next 30 days based on 6 months' historical data
  • Forecast daily temperature for the next 10 days based on 3 months' historical data

Tip: Still curious? You can find an in-depth explanation of time series forecasting in this article.