It’s a great day in sports analytics! Our topic of discussion for today is supervised and unsupervised learning. Within machine learning, there are two main types of tasks: supervised and unsupervised. Today, we’ll learn what is supervised learning and how it relates to unsupervised learning.
Coaches and sports analysts use the data mining process of supervised and unsupervised learning techniques to group players or athletes in a team or a team in a sport with specific goals, such as finding structure in sports training data. With this method, coaches hope to design better structured training sessions and game plans, which will eventually increase the performance of the team or the athlete under tutelage.
Supervised Learning as a Method of Sport Analytics
Supervised learning is an algorithm that is used to analyze the training data and provides a function that can then be used for mapping out new examples other than the initial input. This function reveals the scenario with the best result that can be obtained from an unseen event. Let’s look at the following example:
Suppose you’re a coach that trains different players or athletes. You can group your players according to a particular theme or category such as height range, position, power, and so on. You can then teach the machine using the factors of categorizing based on their specifications. These factors become the data you feed the machine with. In this case, because of your presence to instruct the machine, it is called “supervised learning”.
Now, the machine learns from those factors you input. Each time you specify the characteristics of another player, the machine immediately categorizes the player. This may be done for several purposes including enhancing the training drills you give your players.
Suppose you instruct the machine to keep the data according to drills. The performance of each player across these drills can be averaged and stored. This becomes the data set of which the amount reflects in the number of observations. For instance, if the average drills stored gives a data set of 10, the expected number of observations should correspond to each of the players. These observations can help a coach determine which drills work best for which players.
There are many supervised learning procedures such as linear regression, decision trees, logistic regression, random forest, etc.
Unsupervised Learning as a Method of Sport Analytics
Now, remember that in the case of supervised learning, the coach or analyst is there to instruct the machine with certain factors (or classifications) which are converted into data. But under the unsupervised learning technique, there are no classifications.
In unsupervised learning methods, the coach or sports analyst trains the machine by inputting information, which is not categorized or classified. In this case, the machine algorithm acts on its own by searching for similarities, patterns, differences without any prior knowledge of the players, athletes or teams.
Coaches and sports analysts mostly use unsupervised learning techniques to seek out that which is hidden in the structure of drills or performances of the team, players or athletes.
The two basic methods the machine uses under unsupervised learning for sport analytics is clustering and association.
Conclusively, supervised and unsupervised learning methods of sports analytics are advantageous in that they do not require much brainstorming from the coaches or analysts. And yet they give optimum results, which assists coaches in analyzing player performance.
Find out how Sports Analytics Expert Victor Holman can give your team the competitive advantage.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Learn about the Groundbreaking Sports Analytics Model coaches and sports analysts are talking about!
Learn all about Sports Analytics here.