Introduction of fluorescence-based Real-Time PCR (RT-PCR) is increasingly used to detect multiple pathogens simultaneously and rapidly by gene expression analysis of PCR amplification data. PCR data is analyzed often by setting an arbitrary threshold that intersect the signal curve in its exponential phase if it exists. The point at which the curve crosses the threshold is called Threshold Cycle (CT) for positive samples. On the other, when such cross of threshold does not occur, the sample is identified as negative. This simple and arbitrary however not an elagant definition of CT value sometimes leads to conclusions that are either false positive or negative. Therefore, the purpose of this paper is to present a stable and consistent alternative approach that is based on machine learning for the definition and determination of CT values.
{"title":"Machine Learning for Optimum CT-Prediction for qPCR","authors":"M. Günay, Evgin Göçeri, R. Balasubramaniyan","doi":"10.1109/ICMLA.2016.0103","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0103","url":null,"abstract":"Introduction of fluorescence-based Real-Time PCR (RT-PCR) is increasingly used to detect multiple pathogens simultaneously and rapidly by gene expression analysis of PCR amplification data. PCR data is analyzed often by setting an arbitrary threshold that intersect the signal curve in its exponential phase if it exists. The point at which the curve crosses the threshold is called Threshold Cycle (CT) for positive samples. On the other, when such cross of threshold does not occur, the sample is identified as negative. This simple and arbitrary however not an elagant definition of CT value sometimes leads to conclusions that are either false positive or negative. Therefore, the purpose of this paper is to present a stable and consistent alternative approach that is based on machine learning for the definition and determination of CT values.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"129 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130920061","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}
Wireless content delivery networks (WCDNs) have received attention as a promising solution to reduce the network congestion caused by rapidly growing demands for mobile content. The amount of reduced congestion is intuitively proportional to the hit ratio in a WCDN. Cooperation among cache servers is strongly required to maximize the hit ratio in a WCDN where each cache server is equipped with a small-size cache storage space. In this paper, we address a content replacement problem that deals with how to manage contents in a limited cache storage space in a reactive manner to cope with a dynamic content demand over time. As a new challenge, we apply reinforcement learning, which is Q-learning, to the content replacement problem in a WCDN with coooperative caching. We model the content replacement problem as a Markov Decision Process (MDP) and finally propose an efficient content replacement strategy to maximize the hit ratio based on a multi-agent Q-learning scheme. Simulation results exhibit that the proposed strategy contributes to achieving better content delivery performance in delay due to a higher hit ratio, compared to typical existing schemes of least recently used (LRU) and least frequently used (LFU).
{"title":"Efficient Content Replacement in Wireless Content Delivery Network with Cooperative Caching","authors":"Jihoon Sung, Kyounghye Kim, Junhyuk Kim, J. Rhee","doi":"10.1109/ICMLA.2016.0096","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0096","url":null,"abstract":"Wireless content delivery networks (WCDNs) have received attention as a promising solution to reduce the network congestion caused by rapidly growing demands for mobile content. The amount of reduced congestion is intuitively proportional to the hit ratio in a WCDN. Cooperation among cache servers is strongly required to maximize the hit ratio in a WCDN where each cache server is equipped with a small-size cache storage space. In this paper, we address a content replacement problem that deals with how to manage contents in a limited cache storage space in a reactive manner to cope with a dynamic content demand over time. As a new challenge, we apply reinforcement learning, which is Q-learning, to the content replacement problem in a WCDN with coooperative caching. We model the content replacement problem as a Markov Decision Process (MDP) and finally propose an efficient content replacement strategy to maximize the hit ratio based on a multi-agent Q-learning scheme. Simulation results exhibit that the proposed strategy contributes to achieving better content delivery performance in delay due to a higher hit ratio, compared to typical existing schemes of least recently used (LRU) and least frequently used (LFU).","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127862318","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}
The paper presents an unsupervised nonlinear dimensionality reduction algorithm called Unit Ball Embedding (UBE). Many high-dimensional data, such as object or face images, lie on a union of low-dimensional subspaces which are often called manifolds. The proposed method is able to learn the structure of these manifolds by exploiting the local neighborhood arrangement around each point. It tries to preserve the local structure by minimizing a cost function that measures the discrepancy between similarities of points in the high-dimensional data and similarities of points in the low-dimensional embedding. The cost function is proposed in a way that it provides a hyper-spherical representation of points in the low-dimensional embedding. Visualizations of our method on different datasets show that it creates large gaps between the manifolds and maximizes the separability of them. As a result, it notably improves the quality of unsupervised machine learning tasks (e.g. clustering). UBE is successfully applied on image datasets such as faces, handwritten digits, and objects and the results of clustering on the low-dimensional embedding show significant improvement over existing dimensionality reduction methods.
{"title":"Nonlinear Dimensionality Reduction by Unit Ball Embedding (UBE) and Its Application to Image Clustering","authors":"Behrouz Haji Soleimani, S. Matwin","doi":"10.1109/ICMLA.2016.0177","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0177","url":null,"abstract":"The paper presents an unsupervised nonlinear dimensionality reduction algorithm called Unit Ball Embedding (UBE). Many high-dimensional data, such as object or face images, lie on a union of low-dimensional subspaces which are often called manifolds. The proposed method is able to learn the structure of these manifolds by exploiting the local neighborhood arrangement around each point. It tries to preserve the local structure by minimizing a cost function that measures the discrepancy between similarities of points in the high-dimensional data and similarities of points in the low-dimensional embedding. The cost function is proposed in a way that it provides a hyper-spherical representation of points in the low-dimensional embedding. Visualizations of our method on different datasets show that it creates large gaps between the manifolds and maximizes the separability of them. As a result, it notably improves the quality of unsupervised machine learning tasks (e.g. clustering). UBE is successfully applied on image datasets such as faces, handwritten digits, and objects and the results of clustering on the low-dimensional embedding show significant improvement over existing dimensionality reduction methods.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128993519","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}
S. Hayashi, S. Ono, S. Hosoda, M. Numao, Ken-ichi Fukui
In the ocean around the world, depth series ocean data of temperature and salinity have being measured. However, it is difficult to discriminate the errors from the normal data since the variation of ocean areas are different. In this research, using hierarchical clustering, we partitioned the ocean into some areas so that the ocean data have the same variation in each area. Then, transforming the ocean data into sets of sliding windows in consideration of depth series, we applied some anomaly detection methodologies. Finally, we succeeded in assigning high anomaly scores on errors that seemed to be normal.
{"title":"Error Detection of Ocean Depth Series Data with Area Partitioning and Using Sliding Window","authors":"S. Hayashi, S. Ono, S. Hosoda, M. Numao, Ken-ichi Fukui","doi":"10.1109/ICMLA.2016.0186","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0186","url":null,"abstract":"In the ocean around the world, depth series ocean data of temperature and salinity have being measured. However, it is difficult to discriminate the errors from the normal data since the variation of ocean areas are different. In this research, using hierarchical clustering, we partitioned the ocean into some areas so that the ocean data have the same variation in each area. Then, transforming the ocean data into sets of sliding windows in consideration of depth series, we applied some anomaly detection methodologies. Finally, we succeeded in assigning high anomaly scores on errors that seemed to be normal.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126412055","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}
Estimation of worker reliability on microtask crowdsourcing platforms has gained attention from many researchers. On microtask platforms no worker is fully reliable for a task and it is likely that some workers are spammers, in the sense that they provide a random answer to collect the financial reward. Existence of spammers is harmful as they increase the cost of microtasking and will negatively affect the answer aggregation process. Hence, to discriminate spammers and non-spammers one needs to measure worker reliability to predict how likely that a worker put an effort in solving a task. In this paper we introduce a new reliability estimation algorithm works based on bee colony algorithm called REBECO. This algorithm relies on Gaussian process model to estimate reliability of workers dynamically. With bees that go in search of pollen, some are more successful than the others. This maps well to our problem, where some workers (i.e., bees) are more successful than other workers for a given task thus, giving rise to a reliability measure. Answer aggregation with respect to worker reliability rates has been considered as a suitable replacement for conventional majority voting. We compared REBECO with majority voting using two real world datasets. The results indicate that REBECO is able to outperform MV significantly.
{"title":"Bee Colony Based Worker Reliability Estimation Algorithm in Microtask Crowdsourcing","authors":"Alireza Moayedikia, K. Ong, Yee Ling Boo, W. Yeoh","doi":"10.1109/ICMLA.2016.0127","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0127","url":null,"abstract":"Estimation of worker reliability on microtask crowdsourcing platforms has gained attention from many researchers. On microtask platforms no worker is fully reliable for a task and it is likely that some workers are spammers, in the sense that they provide a random answer to collect the financial reward. Existence of spammers is harmful as they increase the cost of microtasking and will negatively affect the answer aggregation process. Hence, to discriminate spammers and non-spammers one needs to measure worker reliability to predict how likely that a worker put an effort in solving a task. In this paper we introduce a new reliability estimation algorithm works based on bee colony algorithm called REBECO. This algorithm relies on Gaussian process model to estimate reliability of workers dynamically. With bees that go in search of pollen, some are more successful than the others. This maps well to our problem, where some workers (i.e., bees) are more successful than other workers for a given task thus, giving rise to a reliability measure. Answer aggregation with respect to worker reliability rates has been considered as a suitable replacement for conventional majority voting. We compared REBECO with majority voting using two real world datasets. The results indicate that REBECO is able to outperform MV significantly.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121186994","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}
Since many sky-survey observations were performed, as well as appreciable amount of data were obtained, study on large-scale evolution of our Universe has become a field of interest. In this work, we concentrate on the X-ray astronomical samples from NASA's Chandra observatory, and propose an approach to classify galaxy clusters (GCs) based on their central gas profiles' morphological features. Firstly, the raw images are preprocessed, and the central gas profile are segmented. Then, the Fourier descriptors and wavelet moments are take advantaged to extract the morphological features. Finally, a tree structure classifier using support vector machine (SVM) is trained and aid us to categorize the X-ray astronomical observations. Experiments and applications of our classification method on the real X-ray astronomical samples were demonstrated, and comparison of our approach with the non-tree SVM classifier was also performed, which proved our approach is accurate and efficient.
{"title":"Classification of X-Ray Galaxy Clusters with Morphological Feature and Tree SVM","authors":"Lei Wang, Zhixian Ma, Haiguang Xu, Jie Zhu","doi":"10.1109/ICMLA.2016.0124","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0124","url":null,"abstract":"Since many sky-survey observations were performed, as well as appreciable amount of data were obtained, study on large-scale evolution of our Universe has become a field of interest. In this work, we concentrate on the X-ray astronomical samples from NASA's Chandra observatory, and propose an approach to classify galaxy clusters (GCs) based on their central gas profiles' morphological features. Firstly, the raw images are preprocessed, and the central gas profile are segmented. Then, the Fourier descriptors and wavelet moments are take advantaged to extract the morphological features. Finally, a tree structure classifier using support vector machine (SVM) is trained and aid us to categorize the X-ray astronomical observations. Experiments and applications of our classification method on the real X-ray astronomical samples were demonstrated, and comparison of our approach with the non-tree SVM classifier was also performed, which proved our approach is accurate and efficient.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"34 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124980238","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}
M. Saadi, Touqeer Ahmad, Yan Zhao, L. Wuttisittikulkij
This paper introduces a novel visible light positioning (VLP) system using an un-supervised machine learning approach. Two transmitters consist of light emitting diodes (LEDs) which are modulated with 1 kHz and 2.5 kHz sinusoidal signals respectively. At the receiver end, the received signal strength (RSS) is calculated and a sparse grid/cube is constructed by measuring light intensity at different locations. A bilinear interpolation is then applied to create a dense grid of readings which is used for the training of a hierarchical k-means clustering system. For a given query LEDs reading, the trained clusters are used for position estimation by minimizing the distances between the readings and cluster centroids. Experimental results show that an average accuracy of 0.31m can be achieved for a room with the dimensions of 4.3 × 4 × 4 m3. We further compared the performance of two other clustering methods: k-medoids and fuzzy c-means however no significant improvement over the kmeans clustering is found.
{"title":"An LED Based Indoor Localization System Using k-Means Clustering","authors":"M. Saadi, Touqeer Ahmad, Yan Zhao, L. Wuttisittikulkij","doi":"10.1109/ICMLA.2016.0048","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0048","url":null,"abstract":"This paper introduces a novel visible light positioning (VLP) system using an un-supervised machine learning approach. Two transmitters consist of light emitting diodes (LEDs) which are modulated with 1 kHz and 2.5 kHz sinusoidal signals respectively. At the receiver end, the received signal strength (RSS) is calculated and a sparse grid/cube is constructed by measuring light intensity at different locations. A bilinear interpolation is then applied to create a dense grid of readings which is used for the training of a hierarchical k-means clustering system. For a given query LEDs reading, the trained clusters are used for position estimation by minimizing the distances between the readings and cluster centroids. Experimental results show that an average accuracy of 0.31m can be achieved for a room with the dimensions of 4.3 × 4 × 4 m3. We further compared the performance of two other clustering methods: k-medoids and fuzzy c-means however no significant improvement over the kmeans clustering is found.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"2012 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128074527","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}
For many years, neural networks have gained gigantic interest and their popularity is likely to continue because of the success stories of deep learning. Nonetheless, their applications are mostly limited to static and not temporal patterns. In this paper, we apply time warping invariant Echo State Networks (ESNs) to time-series classification tasks using datasets from various studies in the UCR archive. We also investigate the influence of ESN architecture and spectral radius of the network in view of general characteristics of data, such as dataset type, number of classes, and amount of training data. We evaluate our results comparing it to other state-of-the-art methods, using One Nearest Neighbor (1-NN) with Euclidean Distance (ED), Dynamic Time Warping (DTW) and best warping window DTW.
{"title":"Time Series Classification Using Time Warping Invariant Echo State Networks","authors":"Pattreeya Tanisaro, G. Heidemann","doi":"10.1109/ICMLA.2016.0149","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0149","url":null,"abstract":"For many years, neural networks have gained gigantic interest and their popularity is likely to continue because of the success stories of deep learning. Nonetheless, their applications are mostly limited to static and not temporal patterns. In this paper, we apply time warping invariant Echo State Networks (ESNs) to time-series classification tasks using datasets from various studies in the UCR archive. We also investigate the influence of ESN architecture and spectral radius of the network in view of general characteristics of data, such as dataset type, number of classes, and amount of training data. We evaluate our results comparing it to other state-of-the-art methods, using One Nearest Neighbor (1-NN) with Euclidean Distance (ED), Dynamic Time Warping (DTW) and best warping window DTW.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"10 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113976588","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}
In this paper, we propose a nonlinear metric learning framework to boost the performance of semi-supervised learning (SSL) algorithms. Constructed on top of Laplacian SVM (LapSVM), the proposed method learns a smooth nonlinear feature space transformation that makes the input data points more linearly separable. Coherent point drifting (CPD) is utilized as the geometric model with the consideration of its remarkable expressive power in generating sophisticated yet smooth deformations. Our framework has broad applicability, and it can be integrated with many other SSL classifiers than LapSVM. Experiments performed on synthetic and real world datasets show the effectiveness of our CPD-LapSVM over the state-of-the-art metric learning solutions in SSL.
{"title":"Nonlinear Metric Learning for Semi-Supervised Learning via Coherent Point Drifting","authors":"P. Zhang, Bibo Shi, Charles D. Smith, Jundong Liu","doi":"10.1109/ICMLA.2016.0058","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0058","url":null,"abstract":"In this paper, we propose a nonlinear metric learning framework to boost the performance of semi-supervised learning (SSL) algorithms. Constructed on top of Laplacian SVM (LapSVM), the proposed method learns a smooth nonlinear feature space transformation that makes the input data points more linearly separable. Coherent point drifting (CPD) is utilized as the geometric model with the consideration of its remarkable expressive power in generating sophisticated yet smooth deformations. Our framework has broad applicability, and it can be integrated with many other SSL classifiers than LapSVM. Experiments performed on synthetic and real world datasets show the effectiveness of our CPD-LapSVM over the state-of-the-art metric learning solutions in SSL.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133934176","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}
R. Hu, K. Leahy, Ioannis C. Konstantakopoulos, D. Auslander, C. Spanos, A. Agogino
Maximising electricity production from wind requires improvement of wind turbine reliability. Component failures result in unscheduled or reactive maintenance on turbines which incurs significant downtime and, in turn, increases production cost, ultimately limiting the competitiveness of renewable energy. Thus, a critical task is the early detection of faults. To this end, we present a framework for fault detection using machine learning that uses Supervisory Control and Data Acquisition (SCADA) data from a large 3MW turbine, supplemented with features derived from this data that encapsulate expert knowledge about wind turbines. These new features are created using application domain knowledge that is general to large horizontal-axis wind turbines, including knowledge of the physical quantities measured by sensors, the approximate locations of the sensors, the time series behaviour of the system, and some statistics related to the interpretation of sensor measurements. We then use mRMR feature selection to select the most important of these features. The new feature set is used to train a support vector machine to detect faults. The classification performance using the new feature set is compared to performance using the original feature set. Use of the new feature set achieves an F1-score of 90%, an improvement of 27% compared to the original feature set.
{"title":"Using Domain Knowledge Features for Wind Turbine Diagnostics","authors":"R. Hu, K. Leahy, Ioannis C. Konstantakopoulos, D. Auslander, C. Spanos, A. Agogino","doi":"10.1109/ICMLA.2016.0056","DOIUrl":"https://doi.org/10.1109/ICMLA.2016.0056","url":null,"abstract":"Maximising electricity production from wind requires improvement of wind turbine reliability. Component failures result in unscheduled or reactive maintenance on turbines which incurs significant downtime and, in turn, increases production cost, ultimately limiting the competitiveness of renewable energy. Thus, a critical task is the early detection of faults. To this end, we present a framework for fault detection using machine learning that uses Supervisory Control and Data Acquisition (SCADA) data from a large 3MW turbine, supplemented with features derived from this data that encapsulate expert knowledge about wind turbines. These new features are created using application domain knowledge that is general to large horizontal-axis wind turbines, including knowledge of the physical quantities measured by sensors, the approximate locations of the sensors, the time series behaviour of the system, and some statistics related to the interpretation of sensor measurements. We then use mRMR feature selection to select the most important of these features. The new feature set is used to train a support vector machine to detect faults. The classification performance using the new feature set is compared to performance using the original feature set. Use of the new feature set achieves an F1-score of 90%, an improvement of 27% compared to the original feature set.","PeriodicalId":356182,"journal":{"name":"2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131892974","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：