Machine Learning

Machine Learning [məˈʃiːn ləːn]: The way computers learn. As in spotify, when they come up with ever better recommendations for you the more you listen to their music.

The most basic idea behind machine learning is that the computer decides on what to do next by looking at past events to determine what worked and what didn’t. This can be done by trial-and-error, much like a child who learns to walk by getting up and falling over and over again. That’s called Reinforcement learning. Or it can look at a given data input and just determine the most similar output, for example when telling if there’s a cat or a dog in your photo.

Mean

Mean [miːn]: a value, often also refered to as the average, which gives some general but quite limited information about a data set.

There are multiple ways to determine the mean of one dataset. Quite common examples are: the median, the arithmetic mean or the truncated mean. Which one you should chose depends on the kind of data in front of you. The main differnce lies in the way these means are affected by extreme values – their so called robustness against anomalies.

Back to Dictionary

Median

Median [ˈmiː.di.ən]: the value of the data-point exactly in the middle of a dataset.

To determine it, the datapoints have to be sorted by value. If you have an odd number of data points, it is the one directly in the middle. With an even number of values you either take the arithmetic mean between the two middle ones or you round up.

Other means are: the arithmetic mean or the truncated medium. Which one you should chose depends on the kind of data in front of you.

The main differnce lies in the way these means are affected by extreme values – their so called robustness against anomalies. The median is very robust against anomalies.

Back to Dictionary