The project is built using Machine Learning as the foundation.
Online gaming and offline gaming are two different aspects that are to be considered while identifying features for our problem. Sometimes, there’s more to a story than what is visible. For instance, we deducted the human nature of instincts and risk-taking capability. These features may play an important role as a signal to generate the required result. For example, if we see that the user is generally a low-risk taker and a bad aimer, and he starts to take unnecessary risks and starts making consistent progress in terms of scores, which might signal our solution to act correspondingly. Based on this, all the features are extracted.
Data collection becomes an important part as we thrive towards achieving the highest accuracy. It can be collected from the gaming company or manually collected. However, the data were obtained while the real gamers were playing the game.
Pre-processing, Feature Engineering, and Model selection
Pre-processing will help in the normalization of data. There are missing values that need to be filled in, to train the model. Our specific data included Risk-Taking capabilities, Income Level, Business Owner/Job Oriented, and alike. The model was selected based on the data. Model improvisation goes on with increasing data and thus, the efficiency keeps on increasing as the model keeps on getting trained.