EEG Signal Classification for Concealed Information Test Using Spider Monkey Candidate Rule Miner.

Controlling continuous data, such as Electroencephalographic (EEG) data, is challenging. The EEG data is recorded during the Concealed Information Test (CIT), performed with a 16-channel electrode. Some of the meta-heuristic algorithms that are already in use need to be improved so that they can handle the balance between the exploration and exploitation stages, deal with the problem of local optimums, and take continuous data. The Spider Monkey Optimization (SMO) algorithm is utilized as a candidate rule miner known as the SM-Candidate Rule Miner (SM-CRM) for categorizing EEG data into two groups, like guilty or innocent. These problems are solved by making an exhaustive optimal rule set that balances the SM-CRM’s sensitivity, accuracy, and specificity. These rules are applied to the entire EEG dataset, and ten cross-folder validations are performed to get the individual confusion matrix. The SM-CRM method was better than other well-known algorithms regarding average sensitivity, specificity, and classification accuracy. It achieved a decent mean rule length and mean rule set size. Compared to meta-heuristic algorithms like the genetic and bat, the proposed framework achieved a maximum accuracy of 97.66