长短期记忆(LSTM)是一种用于深度学习领域的人工递归神经网络(RNN)架构。与标准前馈神经网络不同,LSTM具有反馈连接。它不仅可以处理单个数据点(例如图像),而且可以处理整个数据序列(例如语音或视频)。例如,LSTM适用于诸如未分段的连接手写识别,语音识别和网络流量或IDS(入侵检测系统)中的异常检测之类的任务。常见的LSTM单元由单元,输入门,输出门和忘记门组成。单元会记住任意时间间隔内的值,并且三个门控制着进出单元的信息流。LSTM网络非常适合基于时间序列数据进行分类,处理和做出预测,因为时间序列中重要事件之间可能存在未知持续时间的滞后。开发LSTM是为了解决训练传统RNN时可能遇到的梯度消失问题。与缝隙长度相对不敏感是LSTM在众多应用中优于RNN,隐马尔可夫模型和其他序列学习方法的优势。

最新论文

The recent global outbreak of covid-19 is affecting many countries around the world. Due to the growing number of newly infected individuals and the health-care system bottlenecks, it will be useful to predict the upcoming number of patients. This study aims to efficiently forecast the is used to estimate new cases, number of deaths, and number of recovered patients in Iran for 180 days, using the official dataset of the Iranian Ministry of Health and Medical Education and the impact of control measures on the spread of COVID-19. Four different types of forecasting techniques, time series, and machine learning algorithms, are developed and the best performing method for the given case study is determined. Under the time series, we consider the four algorithms including Prophet, Long short-term memory, Autoregressive, Autoregressive Integrated Moving Average models. On comparing the different techniques, we found that deep learning methods yield better results than time series forecasting algorithms. More specifically, the least value of the error measures is observed in seasonal ANN and LSTM models. Our findings showed that if precautionary measures are taken seriously, the number of new cases and deaths will decrease, and the number of deaths in September 2021 will reach zero.

0
0
下载
预览
Top