In order to determine the validity of this method, they represent the above two hypothesis in the Bayesian Knowledge Tracing model by introducing a novel modification to the model that allows us to fit a same day and new day parameter for one parameter in a conditional probability table (CPT) while keeping the other parameter in the CPT constant. In Knowledge tracing; learn and forget share a CPT and guess and slip share a CPT. The difference between split-KT and the original-KT is the ability to separate the forget, learn, guess, and slip parameters individually. The equivalence between these two KT models was confirmed empirically by learning parameters for each model from a shared dataset, without new day data, and confirming that the learned parameters and predictions were the same. The individualization of the four parameters were achieved by adding a forget node and a learn node to the knowledge node, as well as adding a guess node and slip node to the question node. Therefore, the knowledge nodes and question nodes are conditioned upon the four new nodes. This model can easily set individualized learn rates, forget rates, guess rates and slip rates. By this way the researchers fix the learn parameter and guess parameter in order to investigate how new day instances would affect the forget and slip parameters.