For thorough understanding of the discussion below, refer to the diagram below to have an overview supervised learning. This is necessary to better understanding of machine learning. In mastering machine learning, supervised learning is a mandatory.
If you already have the data and you would like to learn about the future, will you be able to predict your future performance? Is it possible to find the relationship of the past and the present data and use it for a trajectory lucrative performance? Yes, these are all possible. This is where supervised learning sets in.
In supervised learning, you find the mapping between the inputs and the outputs using the values to train a model. It means to say that you already have the data. You use this kind of learning when you want to predict unknown answers from the data you already have. These data are divided into two parts. First is the data that you will use to teach the system or the data set. And second is the data you will use to see if the computer’s algorithms are accurate or the so called test set or hold out data.
For better understanding, let us have this example. Based on previous experiences of the fluctuations of the stock market, you could hard guess the its future movement . From 2001 until 2014, your stock had been showing and increasing trend. However, the onset of 2015 until the first half of 2016, you had been losing an average of .05%. This is what you call data set. Supervised learning looks for pattern of the given values, which is, in this case, the stock market. It can use any information that is relevant like the day of the week, the season, or any political events. If it has already found the pattern, it can now make predictions. Then, this is where hold-out data comes in.
This learning tool is so important to survive in today’s competition. If you would like to have an edge, you need to master it.
Read this article for more information about analytics http://digilitiks.com/google-analytics-101-installing-tracking-tag/