The Pulse of Urban Transport: Exploring the Co-evolving Pattern for Spatio-temporal Forecasting

2021 ◽  
Vol 15 (6) ◽  
pp. 1-25
Author(s):  
Jinliang Deng ◽  
Xiusi Chen ◽  
Zipei Fan ◽  
Renhe Jiang ◽  
Xuan Song ◽  
...  
Keyword(s):  
Time Series Data ◽  
Demand Forecasting ◽  
Intrinsic Property ◽  
Urban Transport ◽  
Series Data ◽  
Transportation Demand ◽  
Demand Information ◽  
Spatio Temporal ◽  
Mode Of Transport ◽  
Pattern Information

Transportation demand forecasting is a topic of large practical value. However, the model that fits the demand of one transportation by only considering the historical data of its own could be vulnerable since random fluctuations could easily impact the modeling. On the other hand, common factors like time and region attribute, drive the evolution demand of different transportation, leading to a co-evolving intrinsic property between different kinds of transportation. In this work, we focus on exploring the co-evolution between different modes of transport, e.g., taxi demand and shared-bike demand. Two significant challenges impede the discovery of the co-evolving pattern: (1) diversity of the co-evolving correlation, which varies from region to region and time to time. (2) Multi-modal data fusion. Taxi demand and shared-bike demand are time-series data, which have different representations with the external factors. Moreover, the distribution of taxi demand and bike demand are not identical. To overcome these challenges, we propose a novel method, known as co-evolving spatial temporal neural network (CEST). CEST learns a multi-view demand representation for each mode of transport, extracts the co-evolving pattern, then predicts the demand for the target transportation based on multi-scale representation, which includes fine-scale demand information and coarse-scale pattern information. We conduct extensive experiments to validate the superiority of our model over the state-of-art models.

scholarly journals Tourism Demand Forecasting Based on an LSTM Network and Its Variants

Algorithms ◽  
2021 ◽  
Vol 14 (8) ◽  
pp. 243
Author(s):  
Shun-Chieh Hsieh
Keyword(s):  
Adaptive Learning ◽  
Time Series Data ◽  
Short Term Memory ◽  
Demand Forecasting ◽  
Series Data ◽  
Learning Approaches ◽  
Tourism Demand ◽  
Long Short Term Memory ◽  
Lstm Network ◽  
Gated Recurrent Unit

The need for accurate tourism demand forecasting is widely recognized. The unreliability of traditional methods makes tourism demand forecasting still challenging. Using deep learning approaches, this study aims to adapt Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), and Gated Recurrent Unit networks (GRU), which are straightforward and efficient, to improve Taiwan’s tourism demand forecasting. The networks are able to seize the dependence of visitor arrival time series data. The Adam optimization algorithm with adaptive learning rate is used to optimize the basic setup of the models. The results show that the proposed models outperform previous studies undertaken during the Severe Acute Respiratory Syndrome (SARS) events of 2002–2003. This article also examines the effects of the current COVID-19 outbreak to tourist arrivals to Taiwan. The results show that the use of the LSTM network and its variants can perform satisfactorily for tourism demand forecasting.

A deep learning method for data recovery in sensor networks using effective spatio-temporal correlation data

Sensor Review ◽  
2019 ◽  
Vol 39 (2) ◽  
pp. 208-217 ◽  
Author(s):  
Jinghan Du ◽  
Haiyan Chen ◽  
Weining Zhang
Keyword(s):  
Deep Learning ◽  
Sensor Networks ◽  
Large Scale ◽  
Time Series Data ◽  
Temporal Correlation ◽  
Data Recovery ◽  
Series Data ◽  
Learning Method ◽  
Content Type ◽  
Spatio Temporal

Purpose In large-scale monitoring systems, sensors in different locations are deployed to collect massive useful time-series data, which can help in real-time data analytics and its related applications. However, affected by hardware device itself, sensor nodes often fail to work, resulting in a common phenomenon that the collected data are incomplete. The purpose of this study is to predict and recover the missing data in sensor networks. Design/methodology/approach Considering the spatio-temporal correlation of large-scale sensor data, this paper proposes a data recover model in sensor networks based on a deep learning method, i.e. deep belief network (DBN). Specifically, when one sensor fails, the historical time-series data of its own and the real-time data from surrounding sensor nodes, which have high similarity with a failure observed using the proposed similarity filter, are collected first. Then, the high-level feature representation of these spatio-temporal correlation data is extracted by DBN. Moreover, to determine the structure of a DBN model, a reconstruction error-based algorithm is proposed. Finally, the missing data are predicted based on these features by a single-layer neural network. Findings This paper collects a noise data set from an airport monitoring system for experiments. Various comparative experiments show that the proposed algorithms are effective. The proposed data recovery model is compared with several other classical models, and the experimental results prove that the deep learning-based model can not only get a better prediction accuracy but also get a better performance in training time and model robustness. Originality/value A deep learning method is investigated in data recovery task, and it proved to be effective compared with other previous methods. This might provide a practical experience in the application of a deep learning method.

The Digital Earth Viewer: A new visualization approach for geospatial time series data

2021 ◽  
Author(s):  
Valentin Buck ◽  
Flemming Stäbler ◽  
Everardo Gonzalez ◽  
Jens Greinert
Keyword(s):  
Time Series Data ◽  
Environmental Data ◽  
Series Data ◽  
Multiple Sources ◽  
Data Types ◽  
New Approach ◽  
Digital Earth ◽  
Spatio Temporal ◽  
Diverse Data

<p>The study of the earth’s systems depends on a large amount of observations from homogeneous sources, which are usually scattered around time and space and are tightly intercorrelated to each other. The understanding of said systems depends on the ability to access diverse data types and contextualize them in a global setting suitable for their exploration. While the collection of environmental data has seen an enormous increase over the last couple of decades, the development of software solutions necessary to integrate observations across disciplines seems to be lagging behind. To deal with this issue, we developed the Digital Earth Viewer: a new program to access, combine, and display geospatial data from multiple sources over time.</p><p>Choosing a new approach, the software displays space in true 3D and treats time and time ranges as true dimensions. This allows users to navigate observations across spatio-temporal scales and combine data sources with each other as well as with meta-properties such as quality flags. In this way, the Digital Earth Viewer supports the generation of insight from data and the identification of observational gaps across compartments.</p><p>Developed as a hybrid application, it may be used both in-situ as a local installation to explore and contextualize new data, as well as in a hosted context to present curated data to a wider audience.</p><p>In this work, we present this software to the community, show its strengths and weaknesses, give insight into the development process and talk about extending and adapting the software to custom usecases.</p>

scholarly journals A Hybrid Deep Learning Framework for Unsupervised Anomaly Detection in Multivariate Spatio-Temporal Data

2020 ◽  
Vol 10 (15) ◽  
pp. 5191
Author(s):  
Yıldız Karadayı ◽  
Mehmet N. Aydin ◽  
A. Selçuk Öğrenci
Keyword(s):  
Deep Learning ◽  
Anomaly Detection ◽  
Hurricane Katrina ◽  
Time Series Data ◽  
Hybrid Approach ◽  
Ground Truth ◽  
Outbreak Detection ◽  
Series Data ◽  
Detection Techniques ◽  
Spatio Temporal

Multivariate time-series data with a contextual spatial attribute have extensive use for finding anomalous patterns in a wide variety of application domains such as earth science, hurricane tracking, fraud, and disease outbreak detection. In most settings, spatial context is often expressed in terms of ZIP code or region coordinates such as latitude and longitude. However, traditional anomaly detection techniques cannot handle more than one contextual attribute in a unified way. In this paper, a new hybrid approach based on deep learning is proposed to solve the anomaly detection problem in multivariate spatio-temporal dataset. It works under the assumption that no prior knowledge about the dataset and anomalies are available. The architecture of the proposed hybrid framework is based on an autoencoder scheme, and it is more efficient in extracting features from the spatio-temporal multivariate datasets compared to the traditional spatio-temporal anomaly detection techniques. We conducted extensive experiments using buoy data of 2005 from National Data Buoy Center and Hurricane Katrina as ground truth. Experiments demonstrate that the proposed model achieves more than 10% improvement in accuracy over the methods used in the comparison where our model jointly processes the spatial and temporal dimensions of the contextual data to extract features for anomaly detection.

A New Imputation Method for Treating Missing Precipitation Records

2014 ◽  
Vol 635-637 ◽  
pp. 1488-1495
Author(s):  
Yu Liu ◽  
Feng Rui Chen
Keyword(s):  
Time Series Data ◽  
Absolute Error ◽  
Imputation Method ◽  
Temporal Information ◽  
Series Data ◽  
Precipitation Trend ◽  
Monthly Total Precipitation ◽  
The Mean ◽  
Spatio Temporal ◽  
Precipitation Records

This study aims to present a new imputation method for missing precipitation records by fusing its spatio-temporal information. On the basis of extending simple kriging model, a nonstationary kriging method which assumes that the mean or trend is known and varies in whole study area was proposed. It obtains precipitation trend of each station at a given time by analyzing its time series data, and then performs geostatistical analysis on the residual between the trend and measured values. Finally, these spatio-temporal information is integrated into a unified imputation model. This method was illustrated using monthly total precipitation data from 671 meteorological stations of China in April, spanning the period of 2001-2010. Four different methods, including moving average, mean ratio, expectation maximization and ordinary kriging were introduced to compare with. The results show that: Among these methods, the mean absolute error, mean relative error and root mean square error of the proposed method are the smallest, so it produces the best imputation result. That is because: (1) It fully takes into account the spatio-temporal information of precipitation. (2) It assumes that the mean varies in whole study area, which is more in line with the actual situation for rainfall.

scholarly journals Uncertainty Analysis of Object-Based Land-Cover Classification Using Sentinel-2 Time-Series Data

2020 ◽  
Vol 12 (22) ◽  
pp. 3798
Author(s):  
Lei Ma ◽  
Michael Schmitt ◽  
Xiaoxiang Zhu
Keyword(s):  
Time Series ◽  
Land Cover ◽  
Time Series Data ◽  
Land Cover Classification ◽  
Series Data ◽  
Spectral Bands ◽  
Object Based ◽  
Spatio Temporal ◽  
Optical Satellite Images ◽  
The Impact

Recently, time-series from optical satellite data have been frequently used in object-based land-cover classification. This poses a significant challenge to object-based image analysis (OBIA) owing to the presence of complex spatio-temporal information in the time-series data. This study evaluates object-based land-cover classification in the northern suburbs of Munich using time-series from optical Sentinel data. Using a random forest classifier as the backbone, experiments were designed to analyze the impact of the segmentation scale, features (including spectral and temporal features), categories, frequency, and acquisition timing of optical satellite images. Based on our analyses, the following findings are reported: (1) Optical Sentinel images acquired over four seasons can make a significant contribution to the classification of agricultural areas, even though this contribution varies between spectral bands for the same period. (2) The use of time-series data alleviates the issue of identifying the “optimal” segmentation scale. The finding of this study can provide a more comprehensive understanding of the effects of classification uncertainty on object-based dense multi-temporal image classification.

Autoregressive-based outlier algorithm to detect money laundering activities

2017 ◽  
Vol 20 (2) ◽  
pp. 190-202 ◽  
Author(s):  
Kannan S. ◽  
Somasundaram K.
Keyword(s):  
Time Series ◽  
Outlier Detection ◽  
Money Laundering ◽  
Time Series Data ◽  
Demand Forecasting ◽  
Ar Model ◽  
Series Data ◽  
Content Type ◽  
Time Consumption ◽  
Time Series Mining

Purpose Due to the large-size, non-uniform transactions per day, the money laundering detection (MLD) is a time-consuming and difficult process. The major purpose of the proposed auto-regressive (AR) outlier-based MLD (AROMLD) is to reduce the time consumption for handling large-sized non-uniform transactions. Design/methodology/approach The AR-based outlier design produces consistent asymptotic distributed results that enhance the demand-forecasting abilities. Besides, the inter-quartile range (IQR) formulations proposed in this paper support the detailed analysis of time-series data pairs. Findings The prediction of high-dimensionality and the difficulties in the relationship/difference between the data pairs makes the time-series mining as a complex task. The presence of domain invariance in time-series mining initiates the regressive formulation for outlier detection. The deep analysis of time-varying process and the demand of forecasting combine the AR and the IQR formulations for an effective outlier detection. Research limitations/implications The present research focuses on the detection of an outlier in the previous financial transaction, by using the AR model. Prediction of the possibility of an outlier in future transactions remains a major issue. Originality/value The lack of prior segmentation of ML detection suffers from dimensionality. Besides, the absence of boundary to isolate the normal and suspicious transactions induces the limitations. The lack of deep analysis and the time consumption are overwhelmed by using the regression formulation.

scholarly journals Multimodal Data Guided Spatial Feature Fusion and Grouping Strategy for E-Commerce Commodity Demand Forecasting

2021 ◽  
Vol 2021 ◽  
pp. 1-14
Author(s):  
Weiwei Cai ◽  
Yaping Song ◽  
Zhanguo Wei
Keyword(s):  
Time Series Data ◽  
Short Term Memory ◽  
Feature Fusion ◽  
Demand Forecasting ◽  
Spatial Relations ◽  
Series Data ◽  
Multimodal Data ◽  
Spatial Feature ◽  
Grouping Strategy ◽  
Commodity Demand

E-commerce offers various merchandise for selling and purchasing with frequent transactions and commodity flows. An accurate prediction of customer needs and optimized allocation of goods is required for cost reduction. The existing solutions have significant errors and are unsuitable for addressing warehouse needs and allocation. That is why businesses cannot respond to customer demands promptly, as they need accurate and reliable demand forecasting. Therefore, this paper proposes spatial feature fusion and grouping strategies based on multimodal data and builds a neural network prediction model for e-commodity demand. The designed model extracts order sequence features, consumer emotional features, and facial value features from multimodal data from e-commerce products. Then, a bidirectional long short-term memory network- (BiLSTM-) based grouping strategy is proposed. The proposed strategy fully learns the contextual semantics of time series data while reducing the influence of other features on the group’s local features. The output features of multimodal data are highly spatially correlated, and this paper employs the spatial dimension fusion strategy for feature fusion. This strategy effectively obtains the deep spatial relations among multimodal data by integrating the features of each column in each group across spatial dimensions. Finally, the proposed model’s prediction effect is tested using e-commerce dataset. The experimental results demonstrate the proposed algorithm’s effectiveness and superiority.

Sign in / Sign up

Export Citation Format

Share Document