文章基本信息

标题：Predictive Analysis of RFID Supply Chain Path Using Long Short Term Memory (LSTM): Recurrent Neural Networks
本地全文：下载
作者：Meghna Sharma ; Manjeet Singh Tomer
期刊名称：International Journal of Wireless and Microwave Technologies(IJWMT)
印刷版ISSN：2076-1449
电子版ISSN：2076-9539
出版年度：2018
卷号：8
期号：4
页码：66-77
DOI：10.5815/ijwmt.2018.04.05
出版社：MECS Publisher
摘要：Prediction of location has gained lot of attention in different applications areas like predicting the path or any deviation like taxi-route, bus route, human trajectory, robot navigation. Prediction of the next location or any path deviation in RFID enabled supply chain path followed in the process is quite a novel area for the related techniques. The paper defines the architecture for the outlier detection in RFID enabled Supply Chain Path based on historical datasets .Given the training datasets, different classification models are compared for the accurate prediction of the outlierness of the path followed by the tagged objects read by RFID readers during the supply chain process. Comparison of Hidden Markov Model(HMM), XGBoost(decision tree based boosting),Recurrent Neural Network(RNN) and state of art technique in RNN known as Long Short Term Memory (LSTM) is done .To our knowledge LSTM has never been used for this application area for outlier prediction. For the longer path sequences, LSTM has outperformed over other techniques. The training datasets used here are in the form of the record of the outlier positions in particular path and at particular time and location.
关键词：Outlier prediction;Recurrent Neural Networks (RNN);Long Short Term Memory (LSTM);Supply Chain;Radio Frequency Identification (RFID)