# iSeizdiag: toward the framework development of epileptic seizure detection for healthcare

## TL;DR

This paper presents a machine learning framework for detecting epileptic seizures using EEG signals with high accuracy.

## Contribution

The novel contribution is an optimized machine learning framework achieving high accuracy for epileptic seizure detection.

## Key findings

- An average accuracy of 95% was achieved during training and validation.
- A 97% accuracy was achieved after testing the optimized model.
- Statistical parameters were calculated to validate the framework's performance.

## Abstract

The seizure episodes result from abnormal and excessive electrical discharges by a group of brain cells. EEG framework-based signal acquisition is the real-time module that records the electrical discharges produced by the brain cells. The electrical discharges are amplified and appear as a graph on electroencephalogram systems. Different neurological disorders are represented as different waves on EEG records.

This paper involves the detection of Epilepsy which appears as rapid spiking on electroencephalogram signals, using feature extraction and machine learning techniques. Various models, such as the Support Vector Machine, K Nearest Neighbor, and random forest, have been trained, and accuracy has been analyzed to predict the seizure.

An average accuracy of 95% has been claimed using the optimized model for epileptic seizure detection during training and validation. During the analysis of multiple models, the 97% accuracy is claimed after testing. Some statistical parameters are calculated to justify the optimized framework.

The proposed approach represents a satisfactory contribution in precise detection for smart healthcare.