Your own Skyrim - Decision Tree Player Classifier

Your Own Skyrim - Decision Tree Player Classifier is a module that adds a system to Skyrim that records and analyzes player actions and interactions in order to classify players into one of several archetypes. The purpose of this category is to present new customization content to players in the form of skill-focused quests.

Overview

Your Own Skyrim - Decision Tree Player Sorter is a module that adds systems to Skyrim. The system records and analyzes player actions and interactions in order to categorize players into one of several archetypes. The purpose of this category is to present new customization content to players in the form of skill-focused quests.

This was part of my master's thesis project, where I developed and integrated a machine learning algorithm into Skyrim to classify players into one of Richard Bartle's player archetypes based on their behavior and interaction with the virtual world.

This mod is the result of over 170 hours of work in Creation Kit by my supervisor and I, and I really hope you enjoy it. One of the goals of this work is to demonstrate that player modeling systems are a beneficial and popular feature in RPGs, and that overall players will enjoy a personalized experience rather than a generic one.

If you are interested in helping me with my master's thesis, you can fill out this survey, which only takes 3 minutes and means a lot to me.

Thanks

What is it?

The system implemented in this module consists of a decision tree that is run periodically in order to analyze the player's recent actions and assign him a player type and a player category.

The available player types are as follows:

Achievers - Achievers are interested in moving around in the world. They are typical gamers who are looking for "winning". They set themselves game-related goals and are motivated to achieve them.

Explorer - Explorer loves to interact with the world. They rejoiced in their discovery. They try to understand as much as possible about the topology and physics of the environment.

Socializer - Socializers are interested in interacting with other players/NPCS. They spend a lot of time chatting and empathizing with other players and NPCs.

Killer - A killer who acts like other players/NPCS. They want to dominate them through bullying or politicization. They use in-game tools to cause distress to other players/NPCs.

The available classes are as follows:

mage

blacksmith

hunter

How you choose a course depends on your gaming skills. For example, if you have a higher skill level in Destruction and Magic than other skills, you might be classified as a Mage, etc. Player types, on the other hand, use a different approach. These are determined based on your actions during the specified time intervals during which the algorithm collects data. Decision trees take a large number of actions as input, including clearing dungeons, killing people, discovering locations, stealing items, etc.

The classifier was built using a training dataset consisting of 25 individuals under supervised testing. For anyone more interested in this topic, the algorithm used is the C4.5 algorithm.

The algorithm runs every two hours to analyze actions performed and current skills in the previous two hours to update the player's category and type.

After that, the game will adjust itself and select something that may please players of the chosen category and type. This content comes in the form of several class-based tasks developed using the Radiant AI system. This means that certain targets and starting positions for each task will be selected on the fly (while the algorithm is running) and are not predefined. These tasks are also related to player type. For example, if the classifier classifies you as a "killer", you can expect to perform a mission in which you kill multiple enemies. If you are classified as an "explorer," the tasks you will receive may involve getting from point A to point B, etc.

How to use?

Play just like in normal pla …