Optimum analysis of brain activities by using classification and learning techniques
The automated brain activity & stress identification is a crucial issue for environmentally-friendly solutions. A illegal day in the brain & stress field are now a big concern for our society, as they are responsible for many health issues and massive losses in enterprises. The problem is rapidly gravitating through constantly high mental workloads and non-stop advancement of technology, leading to continuous change and adaptation needs. Because of the enormous potential for many therapeutic uses, the analysis of brain functions using brain interfacing (BCI) is of recent interest. Particularly for many patients with neurological disabilities and the advancement of technologies appropriate for long term BCI applications. Within medical industries the study of brain functions has become a recurrent topic because of its growing importance within modern society. The key explanation behind these situations is to consider their tension in human life.
Often stress is characterized as a reaction of the body to perceived physical, mental or emotional distress. A system that is fully automated is desirable in order to solve and fix these problems. With this motivation, this work aims to design a framework for stress recognition system with more accuracy and reliability to recognize real-time stress based on Electroencephalography (EEG) signals. By using Different frequency band & principal component analysis (kernel PCA) feature extraction method it gives the influential features for stress recognition. This process is used to transform the EEG signal into more informative signal that can be considered as a sample dataset. It gives the extracted information from that sample dataset to achieve the intended stress of person. In this research, with three EEG frequency bands with machine learning algorithm have been used for feature extraction in order to extract significant data to achieve better classification accuracy of stress recognition.