F Measure in Machine Learning: Everything You Need to Know And Learn About

There are many pieces and things combined to make a perfectly working machine learning model and what most people usually forget is “F Score or F Measure”. This is a very important component of the whole process and you need to understand both its working as well as its usage to get the best working model. In this blog, we will be going to discuss about F measure in machine learning and other aspects of it.

What is an F Measure or F Score and Where Is It Used?

F score or F measure is basically the happy medium between precision and recall. In simpler terms, it tells you how well you got all of the answers and how many of those answers you got right. You might not know, but f-measure is very common in both machine learning as well as analytics across the board.

You can use this in many situations such as the identification of diseases in patients, search engine optimization, cyber security threat detection, supply chains, text classification, and many others. In general, f measure in machine learning is there to help you understand how confident you should be in the model that you have created.

Also, it can tell you how well are those predictions going to work in real life scenarios. When you’re doing machine learning you want to have a training set. This training set could be a predefined training set, something that has been trained on a massive amount of information like neural networks. This is because machines learn from observation whether that’s through text, voice, or images.

F Measure Mathematical Equation

Let’s imagine that you made a machine learning model about predicting whether a text message is spam or not. There are mainly two points you need to focus on. The first is “Precision” and the other is “Recall”. F Measure merges both of them in a single score and tells you how well your machine learning model is doing in classification tasks.

F1 = 2 x (Precision x Recall / Precision + Recall)

What Are Precision and Recall?

To better understand the whole F Measure in Machine Learning you need to learn about Precision as well as Recall. Therefore, let’s have a look at both of them.


“Precision” is represented as True Positive divided by True Positive plus False Positive. From the example above, it tells us how many of the messages are spam. If you are getting a high value that means your model is working effectively and is not labeling non-spam messages as spam.

From a math perspective what it’s saying is that as the number of false positives increases the rate of precision decreases because the dividend grows.

Let’s look at an example let’s say you have 10 true pulse positives and two false positives. In that case, you would have 10 true positives divided by 10 plus 2 which would equal a rate of 83 percent. It means you are dealing with a low number of false positives, which is good in terms of your machine-learning model precision.


“Recall” is represented as True Positive divided by True Positive plus False Negatives. Let us once again look at the above example to understand this equation. Recall tells us how many of the spam flagged messages are actually spam. If you are getting high recall then your spam detection machine learning model is not leaving any spam message un-flagged.

Use Case of F Measure

Before moving on to the use case of F Measure, you need to understand everything up until now. If you created a machine learning model and that model has a high precision value but low recall then that means it is not performing as intended and missing many of the tasks. Similarly, if your model has a very high recall value but low precision then your model is not effective and is actually making things that it is not supposed to do.

You can use F Measure in a lot of things such as Classification Task. The example which we discussed above is a classification model. Besides that, you can use it in “Information Retrieval” which means, you can find out the performance of a search engine or other information retrieval systems. Last but not least, it can be used in “Model Comparision”. It can find the difference in the performance of machine learning models doing the same thing.


F Measure in Machine Learning is a very important aspect and you can use it to find the perfect balance of two factors. In a machine learning model with a high F Score or F Measure, there is a respectable stability between precision and recall. If you want to know more about it in further detail, read the blog above.

Leave a Comment