With classification, you predict categories while in regression, and you generally predict values.
In supervised learning, classification is multi-dimensional in the sense that sometimes you only have two classes (“yes” or “no”, or, “true” or “false”). But, sometimes you have more than two. For instance, under risk management or risk modeling, you can have “low risk”, “medium risk”, or “high risk.” SVM is a binary classifier (a classifier used for those true/false, yes/no types of classification problems).
Features are important in supervised learning. If there are several features, SVM may be the better classification algorithm choice as opposed to logistic regression. Under supervised learning, you present the computer with example inputs and their desired outputs (those known outcomes). The goal is to learn a general rule that maps inputs to those outputs.
Bug detection, customer churn, stock price prediction (not the value of the stock price, but whether or not it will rise or fall), and weather prediction (sunny/not sunny; rain/no rain) are all examples.
Classification algorithms generally take past data (data for which you have known outcomes), train the model, take new data once the model is trained, ingest it, and create predictions (e.g., is it a truck or is it a car?).