In my statistics classes we cover the subjects of Simple and Multiple Linear Regression - they are techniques for making predictions using regression models. If you know the value of one (independent) variable, you can use a model to predict the value of another (dependent) variable. Often this is visualised on a scatter plot where one variable is plotted against the other. But what if the relationship between the two variables is not linear, for example - what if it is curved? Logarithmic Regression is another type of regression that can be used, and I will be introducing it as a topic in this year's upcoming classes.
In my new video below I show how to perform logarithmic regression in Excel using tree growth data. As trees get older, they grow taller, but the increase in height slows down after a few years to give a curved scatter plot. Can this be used to make predictions? Yes it can - see how:
No comments:
Post a Comment