Supervised Learning
Machine learning algorithms feature supervised learning which lets you work on labeled data. In data science labeled and unlabeled data are very common phenomena. Labeled data means the data that will be used to train the model of learning algorithm has definite features. To understand the concept of supervised learning here is an example —
Say there are house buying price data available in a city. Like for 1000 sq ft, the price was 100k $ and for 1500 sq ft its 120k $. Here we see two feature depending on each other. Here area of the house is related to the pricing of the house. Also, the data is well featured and by the logic, we know how are they related so we call this as labeled data. Now using the supervised learning algorithm we can predict the price of the house for any customs area say 1200 sq ft. What we do here is given the set of values we create a model which best fit to requirements and produces the near to best values. The output of this model will be continuous data in terms of the price of the house for a given area. This problem is called a regression problem. Here important to notice that the input set are real answers.
Another problem that comes in supervised learning is the classification problem, which outputs the discrete values. For example, managing the emails or predicting the email is spam or not can be seen as the classification problem. In this, the output can be only discrete like “ 0 or 1” or “Yes or No”. Also, for example, filtering the email type like personal, advertisement, spam in emails is classification problem where the output is discrete values “1 — personal, 2- advertisement, 3-spam”.