Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00058
David Haley, Ehsan Kamalinejad, Jiaofei Zhong
In this paper, we propose a generalized framework for local clustering based on isoperimetric inequalities. We also demonstrate that contemporary approaches are included in its scope and that it can accommodate data sets of different types, including those with overlapping communities. We then present an efficient, greedy algorithm using the new framework and compare the output of the new algorithm with existing methods.
{"title":"IsoClustering: A Generalized Framework for Local Data Clustering","authors":"David Haley, Ehsan Kamalinejad, Jiaofei Zhong","doi":"10.1109/ICMLA.2019.00058","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00058","url":null,"abstract":"In this paper, we propose a generalized framework for local clustering based on isoperimetric inequalities. We also demonstrate that contemporary approaches are included in its scope and that it can accommodate data sets of different types, including those with overlapping communities. We then present an efficient, greedy algorithm using the new framework and compare the output of the new algorithm with existing methods.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115751500","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00132
Marlu da Silva Santos, M. Ladeira, G. V. Erven, Gladston Luiz da Silva
The scope of modern slavery encompasses human trafficking, forced labor, debt bondage and child labor. This article proposes the use of predictive models to identify the risk of modern slavery in Brazilian cities using real socioeconomic, demographic and rescue operations data. The study uses the embedded technique with Lasso regularization (L1) to select variables. A comparative analyze of techniques for treatment of imbalanced data was applied and the results indicated the Random Over-Sampling (ROS) as the best one. In total, 16 models are evaluated, consisting of 8 different data sets and two classifiers: Logistic Regression (LR) and Gradient Boosting Machine (GBM). The results indicate that the GBM model has better performance and efficiency, with accuracy of 77%, AUC 80% and G-mean of 71%.
{"title":"Machine Learning Models to Identify the Risk of Modern Slavery in Brazilian Cities","authors":"Marlu da Silva Santos, M. Ladeira, G. V. Erven, Gladston Luiz da Silva","doi":"10.1109/ICMLA.2019.00132","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00132","url":null,"abstract":"The scope of modern slavery encompasses human trafficking, forced labor, debt bondage and child labor. This article proposes the use of predictive models to identify the risk of modern slavery in Brazilian cities using real socioeconomic, demographic and rescue operations data. The study uses the embedded technique with Lasso regularization (L1) to select variables. A comparative analyze of techniques for treatment of imbalanced data was applied and the results indicated the Random Over-Sampling (ROS) as the best one. In total, 16 models are evaluated, consisting of 8 different data sets and two classifiers: Logistic Regression (LR) and Gradient Boosting Machine (GBM). The results indicate that the GBM model has better performance and efficiency, with accuracy of 77%, AUC 80% and G-mean of 71%.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120982635","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00100
Z. Jiao, J. Oh
In end-to-end reinforcement learning, an agent captures the entire mapping from its raw sensor data to actuation commands using a single neural network. End-to-end reinforcement learning is mostly studied in single-agent domains, and its scalability to multi-agent setting is under-explored. Without effective techniques, learning effective policies based on the joint observation of agents can be intractable, particularly when sensor data perceived by each agent is high-dimensional. Extending the multi-agent actor-critic method MADDPG, this paper presents Rec-MADDPG, an end-to-end reinforcement learning method for multi-agent continuous control in a cooperative environment. To ease end-to-end learning in a multi-agent setting, we proposed two embedding mechanisms, joint and independent embedding, to project agents' joint sensor observation to low-dimensional features. For training efficiency, we applied parameter sharing and the A3C-based asynchronous framework to Rec-MADDPG. Considering the challenges that can arise in real-world multi-agent control, we evaluated Rec-MADDPG in robotic navigation tasks based on realistic simulated robots and physics enable environments. Through extensive evaluation, we demonstrated that Rec-MADDPG can significantly outperform MADDPG and was able to learn individual end-to-end policies for continuous control based on raw sensor data. In addition, compared to joint embedding, independent embedding enabled Rec-MADDPG to learn even better optimal policies.
{"title":"End-to-End Reinforcement Learning for Multi-agent Continuous Control","authors":"Z. Jiao, J. Oh","doi":"10.1109/ICMLA.2019.00100","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00100","url":null,"abstract":"In end-to-end reinforcement learning, an agent captures the entire mapping from its raw sensor data to actuation commands using a single neural network. End-to-end reinforcement learning is mostly studied in single-agent domains, and its scalability to multi-agent setting is under-explored. Without effective techniques, learning effective policies based on the joint observation of agents can be intractable, particularly when sensor data perceived by each agent is high-dimensional. Extending the multi-agent actor-critic method MADDPG, this paper presents Rec-MADDPG, an end-to-end reinforcement learning method for multi-agent continuous control in a cooperative environment. To ease end-to-end learning in a multi-agent setting, we proposed two embedding mechanisms, joint and independent embedding, to project agents' joint sensor observation to low-dimensional features. For training efficiency, we applied parameter sharing and the A3C-based asynchronous framework to Rec-MADDPG. Considering the challenges that can arise in real-world multi-agent control, we evaluated Rec-MADDPG in robotic navigation tasks based on realistic simulated robots and physics enable environments. Through extensive evaluation, we demonstrated that Rec-MADDPG can significantly outperform MADDPG and was able to learn individual end-to-end policies for continuous control based on raw sensor data. In addition, compared to joint embedding, independent embedding enabled Rec-MADDPG to learn even better optimal policies.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123885115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00185
Francis C. Motta, J. Harer, Nicholas Leiby, F. Marinozzi, Scott Novotney, G. Rocklin, Jed Singer, D. Strickland, M. Vaughn, Christopher J. Tralie, R. Bedini, F. Bini, G. Bini, Hamed Eramian, Marcio Gameiro, S. Haase, Hugh K. Haddox
This paper describes a general pipeline for generating optimal vector representations of topological features of data for use with machine learning algorithms. This pipeline can be viewed as a costly black-box function defined over a complex configuration space, each point of which specifies both how features are generated and how predictive models are trained on those features. We propose using state-of-the-art Bayesian optimization algorithms to inform the choice of topological vectorization hyperparameters while simultaneously choosing learning model parameters. We demonstrate the need for and effectiveness of this pipeline using two difficult biological learning problems, and illustrate the nontrivial interactions between topological feature generation and learning model hyperparameters.
{"title":"Hyperparameter Optimization of Topological Features for Machine Learning Applications","authors":"Francis C. Motta, J. Harer, Nicholas Leiby, F. Marinozzi, Scott Novotney, G. Rocklin, Jed Singer, D. Strickland, M. Vaughn, Christopher J. Tralie, R. Bedini, F. Bini, G. Bini, Hamed Eramian, Marcio Gameiro, S. Haase, Hugh K. Haddox","doi":"10.1109/ICMLA.2019.00185","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00185","url":null,"abstract":"This paper describes a general pipeline for generating optimal vector representations of topological features of data for use with machine learning algorithms. This pipeline can be viewed as a costly black-box function defined over a complex configuration space, each point of which specifies both how features are generated and how predictive models are trained on those features. We propose using state-of-the-art Bayesian optimization algorithms to inform the choice of topological vectorization hyperparameters while simultaneously choosing learning model parameters. We demonstrate the need for and effectiveness of this pipeline using two difficult biological learning problems, and illustrate the nontrivial interactions between topological feature generation and learning model hyperparameters.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125234175","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00078
Ken-ichi Fukui, Junya Tanaka, T. Tomita, M. Numao
In this paper, we focus on a method that integrates a physical model into a neural network. This study proposes a neural network that can predict two components, namely outputs based on a physical model and its model discrepancy. To achieve such a goal, we propose a novel neural network architecture and associated loss functions designed based on a target physical model. The physical model is used as a regularizer of spatial behavior where output from the neural network is used as an intermediate variable. Then, the model discrepancy is defined as its residual to the observation value. We also propose a network architecture which has Shared and Non-Shared networks, and the neural network can be trained by alternate optimization. We constructed the proposed method with wind prediction in the upper troposphere based on thermal wind equations as an example. The experimental results demonstrate that the proposed method can achieve higher predictive accuracy than normal convolutional neural network or using thermal wind equation, also the obtained model discrepancy expresses convergence and divergence of wind vectors.
{"title":"Physics-Guided Neural Network with Model Discrepancy Based on Upper Troposphere Wind Prediction","authors":"Ken-ichi Fukui, Junya Tanaka, T. Tomita, M. Numao","doi":"10.1109/ICMLA.2019.00078","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00078","url":null,"abstract":"In this paper, we focus on a method that integrates a physical model into a neural network. This study proposes a neural network that can predict two components, namely outputs based on a physical model and its model discrepancy. To achieve such a goal, we propose a novel neural network architecture and associated loss functions designed based on a target physical model. The physical model is used as a regularizer of spatial behavior where output from the neural network is used as an intermediate variable. Then, the model discrepancy is defined as its residual to the observation value. We also propose a network architecture which has Shared and Non-Shared networks, and the neural network can be trained by alternate optimization. We constructed the proposed method with wind prediction in the upper troposphere based on thermal wind equations as an example. The experimental results demonstrate that the proposed method can achieve higher predictive accuracy than normal convolutional neural network or using thermal wind equation, also the obtained model discrepancy expresses convergence and divergence of wind vectors.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131538635","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00059
Shokoufeh Monjezi Kouchak, A. Gaffar
Driver distraction is one of the primary causes of fatal car accidents in U.S. Analyzing driver behavior using different types of data including driving data, driver status or a combination of them is an emerging machine learning solution to detect the distraction level and notify the driver. Deep learning methods such as recurrent neural networks outperform other machine learning methods in car safety applications. In this paper, we used time-sequenced driving data that we collected in eight driving contexts to measure the driver distraction level. Our RNN is also capable of detecting the type of behavior that caused distraction. We used the driver interaction with the car infotainment system as the distracting activity. Two types of LSTM networks were used including bidirectional LSTM network and attention network. We compare the performance of these two complex networks to that of the simple LSTM in estimating driver behavior. We show that our attention network outperforms the other two, while adding bidirectional LSTM networks enhanced the training process of simple LSTM network.
{"title":"Using Bidirectional Long Short Term Memory with Attention Layer to Estimate Driver Behavior","authors":"Shokoufeh Monjezi Kouchak, A. Gaffar","doi":"10.1109/ICMLA.2019.00059","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00059","url":null,"abstract":"Driver distraction is one of the primary causes of fatal car accidents in U.S. Analyzing driver behavior using different types of data including driving data, driver status or a combination of them is an emerging machine learning solution to detect the distraction level and notify the driver. Deep learning methods such as recurrent neural networks outperform other machine learning methods in car safety applications. In this paper, we used time-sequenced driving data that we collected in eight driving contexts to measure the driver distraction level. Our RNN is also capable of detecting the type of behavior that caused distraction. We used the driver interaction with the car infotainment system as the distracting activity. Two types of LSTM networks were used including bidirectional LSTM network and attention network. We compare the performance of these two complex networks to that of the simple LSTM in estimating driver behavior. We show that our attention network outperforms the other two, while adding bidirectional LSTM networks enhanced the training process of simple LSTM network.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127739659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00304
Joffrey L. Leevy, T. Khoshgoftaar, Richard A. Bauder, Naeem Seliya
Periodic updating of a machine learning model may become necessary because new data could have a distribution that has drifted significantly over time from the original data distribution, thus impacting the model's usefulness. The primary objective of this paper is to evaluate temporal influence on the maintenance of a predictive model. We investigate the impact of using training data from various year-groupings on a model designed to detect Medicare Part B billing fraud. Training datasets are obtained from year-groupings of 2015, 2014-2015, 2013-2015, and 2012-2015. The test dataset is represented by 2016 data. Our study utilizes five popular learners and five class ratios obtained by Random Undersampling. Using the Area Under the Receiver Operating Characteristic (ROC) Curve as the performance metric, our case study indicates that the Logistic Regression learner yields the highest overall value for the yeargrouping of 2013-2015, with a majority-to-minority ratio of 90:10. For the problem of maintaining predictive models for Medicare fraud, we conclude that a sampled dataset should be chosen over the full dataset and that the largest training dataset (i.e., 2012- 2015) does not always produce the best results. To the best of our knowledge, this is the first big data study that examines the influence of time on the maintenance of machine learning models.
{"title":"The Effect of Time on the Maintenance of a Predictive Model","authors":"Joffrey L. Leevy, T. Khoshgoftaar, Richard A. Bauder, Naeem Seliya","doi":"10.1109/ICMLA.2019.00304","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00304","url":null,"abstract":"Periodic updating of a machine learning model may become necessary because new data could have a distribution that has drifted significantly over time from the original data distribution, thus impacting the model's usefulness. The primary objective of this paper is to evaluate temporal influence on the maintenance of a predictive model. We investigate the impact of using training data from various year-groupings on a model designed to detect Medicare Part B billing fraud. Training datasets are obtained from year-groupings of 2015, 2014-2015, 2013-2015, and 2012-2015. The test dataset is represented by 2016 data. Our study utilizes five popular learners and five class ratios obtained by Random Undersampling. Using the Area Under the Receiver Operating Characteristic (ROC) Curve as the performance metric, our case study indicates that the Logistic Regression learner yields the highest overall value for the yeargrouping of 2013-2015, with a majority-to-minority ratio of 90:10. For the problem of maintaining predictive models for Medicare fraud, we conclude that a sampled dataset should be chosen over the full dataset and that the largest training dataset (i.e., 2012- 2015) does not always produce the best results. To the best of our knowledge, this is the first big data study that examines the influence of time on the maintenance of machine learning models.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127774979","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00129
Scott Wahl, John W. Sheppard, Elizabeth A. Shanahan
An important aspect of social networks is the discovery and partitioning of the complex graphs into dense sub-networks referred to as communities. The goal of such partitioning is to find groups who have similar attributes or behaviors. In the realm of politics, it is possible to group individuals with similar political behavior by analyzing campaign finance records. In this paper we use fuzzy hierarchical spectral clustering to find communities with campaign finance networks. Multiple experiments were performed using varying edge weighting, number and type of communities, as well as analyzing multiple different years of voting data. The results show that using the full hierarchy of community assignments for legislators is highly predictive of voting behavior in the US House of Representatives and Senate.
{"title":"Legislative Vote Prediction using Campaign Donations and Fuzzy Hierarchical Communities","authors":"Scott Wahl, John W. Sheppard, Elizabeth A. Shanahan","doi":"10.1109/ICMLA.2019.00129","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00129","url":null,"abstract":"An important aspect of social networks is the discovery and partitioning of the complex graphs into dense sub-networks referred to as communities. The goal of such partitioning is to find groups who have similar attributes or behaviors. In the realm of politics, it is possible to group individuals with similar political behavior by analyzing campaign finance records. In this paper we use fuzzy hierarchical spectral clustering to find communities with campaign finance networks. Multiple experiments were performed using varying edge weighting, number and type of communities, as well as analyzing multiple different years of voting data. The results show that using the full hierarchy of community assignments for legislators is highly predictive of voting behavior in the US House of Representatives and Senate.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128112129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00269
James Spooner, Madeline Cheah, V. Palade, S. Kanarachos, A. Daneshkhah
The safety of vulnerable road users is of paramount importance as transport moves towards fully automated driving. The richness of real-world data required for testing autonomous vehicles is limited, and furthermore, the available data does not have a fair representation of different scenarios and rare events. This work presents a novel approach for the generation of human pose structures, specifically the type of pose structures that would appear to be in pedestrian scenarios. The results show that the generated pedestrian structures are indistinguishable from the ground truth pose structures when classified using a suitably trained classifier. The paper demonstrates that the Generative Adversarial Network architecture can be used to create realistic new training samples, and, in future, new pedestrian events.
{"title":"Generation of Pedestrian Pose Structures using Generative Adversarial Networks","authors":"James Spooner, Madeline Cheah, V. Palade, S. Kanarachos, A. Daneshkhah","doi":"10.1109/ICMLA.2019.00269","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00269","url":null,"abstract":"The safety of vulnerable road users is of paramount importance as transport moves towards fully automated driving. The richness of real-world data required for testing autonomous vehicles is limited, and furthermore, the available data does not have a fair representation of different scenarios and rare events. This work presents a novel approach for the generation of human pose structures, specifically the type of pose structures that would appear to be in pedestrian scenarios. The results show that the generated pedestrian structures are indistinguishable from the ground truth pose structures when classified using a suitably trained classifier. The paper demonstrates that the Generative Adversarial Network architecture can be used to create realistic new training samples, and, in future, new pedestrian events.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133338470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICMLA.2019.00020
Changwei Hu, Yifan Hu, Sungyong Seo
Multivariate time series are routinely encountered in real-world applications, and in many cases, these time series are strongly correlated. In this paper, we present a deep learning structural time series model which can (i) handle correlated multivariate time series input, and (ii) forecast the targeted temporal sequence by explicitly learning/extracting the trend, seasonality, and event components. The trend is learned via a 1D and 2D temporal CNN and LSTM hierarchical neural net. The CNN-LSTM architecture can (i) seamlessly leverage the dependency among multiple correlated time series in a natural way, (ii) extract the weighted differencing feature for better trend learning, and (iii) memorize the long-term sequential pattern. The seasonality component is approximated via a non-liner function of a set of Fourier terms, and the event components are learned by a simple linear function of regressor encoding the event dates. We compare our model with several state-of-the-art methods through a comprehensive set of experiments on a variety of time series data sets, such as forecasts of Amazon AWS Simple Storage Service (S3) and Elastic Compute Cloud (EC2) billings, and the closing prices for corporate stocks in the same category.
{"title":"A Deep Structural Model for Analyzing Correlated Multivariate Time Series","authors":"Changwei Hu, Yifan Hu, Sungyong Seo","doi":"10.1109/ICMLA.2019.00020","DOIUrl":"https://doi.org/10.1109/ICMLA.2019.00020","url":null,"abstract":"Multivariate time series are routinely encountered in real-world applications, and in many cases, these time series are strongly correlated. In this paper, we present a deep learning structural time series model which can (i) handle correlated multivariate time series input, and (ii) forecast the targeted temporal sequence by explicitly learning/extracting the trend, seasonality, and event components. The trend is learned via a 1D and 2D temporal CNN and LSTM hierarchical neural net. The CNN-LSTM architecture can (i) seamlessly leverage the dependency among multiple correlated time series in a natural way, (ii) extract the weighted differencing feature for better trend learning, and (iii) memorize the long-term sequential pattern. The seasonality component is approximated via a non-liner function of a set of Fourier terms, and the event components are learned by a simple linear function of regressor encoding the event dates. We compare our model with several state-of-the-art methods through a comprehensive set of experiments on a variety of time series data sets, such as forecasts of Amazon AWS Simple Storage Service (S3) and Elastic Compute Cloud (EC2) billings, and the closing prices for corporate stocks in the same category.","PeriodicalId":436714,"journal":{"name":"2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA)","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133714751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：