An implementation of anomaly detection in IOT DTA using a deep (OC-NN) with the long short term memory network (LSTM)

Abstract

An Electrocardiography (ECG) signals are accessed mainly to monitor the health condition of the human heart, and the resulting time series signals are analyzed manually by the medical professionals to detect if there are any anomalies such as arrhythmia. Manual Diagnosis of ECG Signals has been often Prone to Errors. An Electrocardiography (ECG) signals are accessed mainly to monitor the health condition of the human heart, and the resulting time series signals are analyzed manually by the medical professionals to detect if there are any anomalies such as arrhythmia. Manual Diagnosis of ECG Signals has been often Prone to Errors. Past work in Automating the Analysis requires extensive Pre-Processing, which is time-consuming and cumbersome. It takes significant time for Heart Patients in their Precarious Condition. There is a requirement of Computational Analysis, which is fast and efficient. Some of the analysis develops the marked features and design a classifier for discriminating between the healthy ECG signals and those which contains Arrhythmia. This method requires knowledge and relevant data of the various types of Arrhythmia of training the model.However, there can be many different and different new types of Arrhythmia can occur, which previously were not a part of the original training set. Thus, it may be wiser to adopt an anomaly detection approach to analyzing them. In this paper, we are utilizing A deep one class neural network (OC-NN) architecture with the Long Short-Term Memory Network (LSTM) units for developing a predictive model from the healthy ECG signals. The probability distribution of the prediction errors from the models, using the Maximum Likelihood Estimate (MLE) is used for indicating anomalous or non-anomalous behavior. The main advantage of using LSTM networks is that the ECG signals be directly applied to the system without any extensive pre-processing as used by other Detection techniques. No Prior information of abnormal signals makes it worthwhile, as it needs to be trained only on average datza. MIT-BIH Arrhythmia Physionet Database has been used to obtain ECG time series data for both non-anomalous periods and irregular periods. Both the Stateful and homeless Modes of LSTM are projected and enforced. The Results from the Stateful LSTM Model show Precision of 99.36% and a TPR/FPR ratio of 145.39. Results are promising and indicate that the Deep-Stacked Long Short-Term Memory Networks (LSTM) models are feasible for detecting anomalies in ECG signals within a short period.