Churn Prediction with Sequential Data Using Long Short Term Memory

In recent years, long short-term memory (LSTM) has been successfully used to model sequential data of variable length. However, LSTM can still experience difficulty in capturing long-term dependencies. In this work, we tried to alleviate this problem by introducing a dynamic skip connection, which can learn to directly connect two dependent words. Since there is no dependency information in the training data, we propose a novel reinforcement learning-based method to model the dependency relationship and connect dependent words. The proposed model computes the recurrent transition functions based on the skip connections, which provides a dynamic skipping advantage over RNNs that always tackle entire sentences sequentially. Our experimental results on three natural language processing tasks demonstrate that the proposed method can achieve better performance than existing methods. In the number prediction experiment, the proposed model outperformed LSTM with respect to accuracy by nearly 20%.

Download Full-text

Forecasting the pavement icing in traffic zone using long short-term memory

Journal of Intelligent & Fuzzy Systems ◽

10.3233/jifs-219022 ◽

2021 ◽

pp. 1-9

Author(s):

Rao Zhongyang ◽

Feng Chunyuan

Keyword(s):

Traffic Safety ◽

Short Term Memory ◽

Freezing Point ◽

Water Film ◽

Model Parameters ◽

Sequential Data ◽

Short Term ◽

Term Memory ◽

Pavement Temperature ◽

Long Short Term Memory

In winter, driving is a difficult due to road icing. It is one of the most unfavorable weather conditions that endangers traffic safety. We gathered data on pavement temperature, freezing point temperature, friction coefficient, pavement water film thickness, ice content, and pavement condition using sensors. Those data are feed into Long Short-Term Memory (LSTM) to predict road icing in the city. Here the primary issue is forecasting the pavement icing of the traffic zone. Our work is an endeavor to use the deep learning method on LSTM to forecast pavement icing on Ji’nan in China. Those Sequential data of pavement icing can process and memorize by LSTM at a specific time. Finally, research results indicate that the performance of the model is very precise. With LSTM model parameters’ help, the sequential data on road icing prediction can also predict the pavement temperature.

Download Full-text

Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction

Sensors ◽

10.3390/s21217333 ◽

2021 ◽

Vol 21 (21) ◽

pp. 7333

Author(s):

Ricardo Petri Silva ◽

Bruno Bogaz Zarpelão ◽

Alberto Cano ◽

Sylvio Barbon Junior

Keyword(s):

Time Series ◽

Short Term Memory ◽

Machine Learning Algorithms ◽

Detection Methods ◽

Sequential Data ◽

Short Term ◽

Term Memory ◽

Wide Range ◽

Long Short Term Memory ◽

Adf Test

A wide range of applications based on sequential data, named time series, have become increasingly popular in recent years, mainly those based on the Internet of Things (IoT). Several different machine learning algorithms exploit the patterns extracted from sequential data to support multiple tasks. However, this data can suffer from unreliable readings that can lead to low accuracy models due to the low-quality training sets available. Detecting the change point between high representative segments is an important ally to find and thread biased subsequences. By constructing a framework based on the Augmented Dickey-Fuller (ADF) test for data stationarity, two proposals to automatically segment subsequences in a time series were developed. The former proposal, called Change Detector segmentation, relies on change detection methods of data stream mining. The latter, called ADF-based segmentation, is constructed on a new change detector derived from the ADF test only. Experiments over real-file IoT databases and benchmarks showed the improvement provided by our proposals for prediction tasks with traditional Autoregressive integrated moving average (ARIMA) and Deep Learning (Long short-term memory and Temporal Convolutional Networks) methods. Results obtained by the Long short-term memory predictive model reduced the relative prediction error from 1 to 0.67, compared to time series without segmentation.

Download Full-text

Auto-Detection of Human Factor Contents on Social Media Posts Using Word2vec and Long Short-Term Memory (LSTM)

Handbook of Research on the Role of Human Factors in IT Project Management - Advances in IT Personnel and Project Management ◽

10.4018/978-1-7998-1279-1.ch001 ◽

2020 ◽

pp. 1-13

Author(s):

Chika Yinka-Banjo ◽

Gafar Lekan Raji ◽

Ifeanyi Precious Ohalete

Keyword(s):

Social Media ◽

Short Term Memory ◽

Sequential Data ◽

Effective Solution ◽

Learning Models ◽

Short Term ◽

Test Dataset ◽

Term Memory ◽

Poor Academic Performance ◽

Long Short Term Memory

The threat posed by cyberbullying to the mental health in our society cannot be overemphasized. Victims of this menace are reported to have suffered poor academic performance, depression, and suicidal thoughts. There is need to find an efficient and effective solution to this problem within the academic environment. In this research, one of the popular deep learning models—long short-term memory (LSTM)—known for its optimized performance in training sequential data was combined with Word2Vec embedding technique to create a model trained for classifying the content of social media post as containing cyberbullying content or otherwise. The result was observed to have shown improvements in its performance with respect to accuracy in the classification task with over 80% of the test dataset correctly classified as against the existing model with about 74.9% accuracy.

Download Full-text

Bidirectional Long Short-Term Memory Neural Networks for Linear Sum Assignment Problems

Applied Sciences ◽

10.3390/app9173470 ◽

2019 ◽

Vol 9 (17) ◽

pp. 3470

Author(s):

Nguyen Minh-Tuan ◽

Yong-Hwa Kim

Keyword(s):

Neural Network ◽

Deep Learning ◽

Short Term Memory ◽

Sequential Data ◽

Short Term ◽

Data Simulation ◽

Assignment Problems ◽

Term Memory ◽

Long Short Term Memory ◽

Fully Connected

Many resource allocation problems can be modeled as a linear sum assignment problem (LSAP) in wireless communications. Deep learning techniques such as the fully-connected neural network and convolutional neural network have been used to solve the LSAP. We herein propose a new deep learning model based on the bidirectional long short-term memory (BDLSTM) structure for the LSAP. In the proposed method, the LSAP is divided into sequential sub-assignment problems, and BDLSTM extracts the features from sequential data. Simulation results indicate that the proposed BDLSTM is more memory efficient and achieves a higher accuracy than conventional techniques.

Download Full-text