Adjusting The Hard Level on Game by a Prediction for Improving Attraction and Business Value

The computer gaming industry has become more popular among the young generation over the years. It became a sport and According to past studies, people choose computer games over physical games (outdoor games) due to their busy lifestyles which crack social relationships and fear of injuries. There are around 2.2 billion gamers around the world. Out of the assessed 7.6 billion individuals living on earth, as of July 2018, that implies nearly 33% of individuals on this planet are gamers. Out of those 2.2 billion gamers, 1.2 billion of the players play PC game. An online survey was conducted to collect information from random game players and non-game players scattered throughout Sri Lanka to identify the problems faced by them when playing the games and why others do not play the game or are discouraged to play games. 354 participants participated in this survey. In which 127 of the participants were Female and 227 Male. The participation age range was identified as 15 to 49 Years. The research team was able to summarize that 80.04% of the participants play games while 19.06% don't play games. Out of the participants who play games, 9.08% are pro players while the rest of the 78% of players are intermediate players. We identified that the main problems faced by most of the players are as follows: Some missions are too easy and boring or players getting stuck in the middle of the level because the level is too hard or because the mission is too hard to complete. The Popularity of games can be varied according to the players' skill level. Tasks of a game can be boring if it's too easy for a skilled player. On the other hand, if it's too hard for a beginner level player, it may tend to give up the game. The other problem identified was the inability of game character customization as per the preference of 92.07% of gamers who participated in the survey. According to their responses, we were able to understand the problems that they faced while playing the game and what they expect in a game. Therefore, the research team proposes a game that changes the level of hardness/complexity of the level according to the players' mood automatically. In simple terms, the game will `adjust and adopt' based on the user's emotions. The next level will be automatically designed based on the player record of the previous level using an algorithm that takes the emotions as inputs. To achieve this, the gamer's heart pulse rate, facial expression, and speed of using the controls are captured and analysed to obtain the most accurate emotion level of the player using machine learning. The level of emotions is taken along with the parameters such as age, gender, and haptic feedback to adjust the hardness level of the level.It is important to identify the player skill level to understand the user's experience as a gamer. To determine a player's skill level in the background, a keylogger captures the pressed key and the time it takes to press the key. It is also necessary to identify the emotional state of the player to provide a unique user experience to the player and it can be done in two ways which are facial expression and heart-pulse. Facial expression detection results in identifying and predicting the basic emotions of an individual such as happiness, anger, and neutral which will be necessary for the adaptive game, and also the emotional range of a player is predicted. Additionally, age, gender, and face-shape are determined. The dataset is trained using machine learning to detect the facial expression and identify the emotion, age, gender, and face-shape of the individual. The Facial Expression of the player will be captured in real-time from the web camera of the computer and the emotion will be identified automatically as well as the age gender and face-shape. The other method is through Heart pulse, Heart pulse also can be used to detect the emotions of the player. There are 3 types of emotion levels Happy, Fear, Neutral. To do this, the heart rate sensor is used to capture the player's heart rate. According to the output received after training the dataset using machine learning algorithms and the skill level output is used to adjust the level of the game while the gamer is playing the game. To obtain a much more accurate emotional level, the heartbeat level, and the facial emotional level are analysed together. We have created a new model for analysing the above data using linear regression. Using the said model, the gender, age, the skill level of the player as well as the difficulty in the level of the game can be identified. The next level of the game will be designed using the data from the previous level and this process will continue until the player quits the game. A model is trained by using machine learning technology to predict character characteristics (moving speed, health regenerating) of the game character. The difficulty level of the player is taken as input for the training process. The output characteristics are given by an in-game module and adjust the characteristics of the game character until the player quits playing. The AI algorithm performs the process of reflecting the characteristics of the player to the in-game character. Age, gender, and face-shape are used as input data. The algorithm process input data to generate a unique set of variables. The variables manipulate the appearance of the in-game character until the player quits playing. Our game is developed within three months; therefore, it is difficult to predict the emotions or obtain a skill level like in an existing commercialized game because the demo game doesn't evoke any emotions or measure the skill level of the player. So, we have tested our application with the other similar existing games and we were able to obtain an accurate emotion and skill level. Even though the game is designed in a way to change the levels and character abilities based on the inputs obtained, we can't do so in other existing games. Therefore, the player records obtained from the games we tested were taken into a text record and analysed separately. Based on the analysed data we were able to obtain an accuracy level of 68%.In the future, the demo game will be further developed and tested with our system to obtain an emotion. The option of changing the game character based on the player's appearance is 75% successful, in which face-shape and gender were used. In the future, we will be using age as well. In the future a profile can be designed based on skill level, therefore making it possible to adjust the game based on a player's past record.This system provides a solution to problems faced by the game players such as the game being too easy and boring or hard and stressful. This can also prevent health issues that can be caused when a person gets too angry or violent and stress.