Search

In this paper an Online Monitoring System is presented, which enables a universal condition monitoring of electrical household appliances. The approach involves recording the power consumption of the respective device and evaluating the measured data using machine learning methods and tools. Thereby an artificial intelligence learns the normal condition of the examined device and can subsequently perform a monitoring. A k -Means algorithm is used for clustering and a One-class Support Vector Machine for classification. This makes it possible to detect anomalies in the operating modes and behavior of the device during subsequent operation. In addition to the methodology, the design and real application of a prototype is presented, in which a suitable sensor technology and a software with artificial intelligence are implemented. [ABSTRACT FROM AUTHOR]

The usage of Building Automation Systems (BAS) and Energy Management Systems (EMS) is indeed becoming ever more common and sophisticated, and seeking to promote energy savings by integrating new sources of data, such as user preferences, in real-time. This paper reviews the existing systems and proposes an innovation in HVAC systems management: a system that tracks the occupants’ preferences, learns from them, and manages HVAC automatically. Our hypothesis was that by developing a learning system based on feedback acquired through the mobile devices of room occupants to optimize the control of a HVAC system, in order to minimize energy consumption while maximizing average user comfort. A prototype solution is described and evaluated by simulation. We show that ambient intelligent systems can be used to control a building’s EMS, effectively reducing energy consumption while maintaining acceptable comfort levels. Our results indicate that employing a k-means machine learning technique enables the automatic configuration of an HVAC system to reduce energy consumption while keeping the majority of occupants within acceptable comfort levels. The developed prototype provides occupants with feedback on ambient variables on a mobile user interface. © 2017 Elsevier Science. All rights reserved. [ABSTRACT FROM AUTHOR]

Clustering and classifying are overriding techniques in machine learning. Neighborhood construction as a key step in these techniques has been extensively used for modeling local relationships between data samples, and constructing global structures from local information. The goal of the neighborhood construction process is to improve the quality of individual data point categorizing. Many applications such as detecting social network communities, bundling related edges, solving location, and routing problems all indicate the importance of this problem. This paper presents theoretical and practical studies of state-of-the-art methods in the context of neighborhood construction which is resulted in a coherent and comprehensive survey to analyze these methods. To this end, significant algorithms of neighborhood construction have been proposed to analyze data points which are very useful for the community of clustering and classifying practitioners since showing the advantages and disadvantages of each algorithm. All of them will be described and discussed deeply in different aspects, which help to select an appropriate solution for problems. A taxonomy of these algorithms is presented and their differences and some important applications are explained. Finally, the future challenges concerning the title of the present paper are outlined. [ABSTRACT FROM AUTHOR]

Multiple kernel k-means (MKKM) clustering algorithm is widely used in many machine learning and computer vision tasks. This algorithm improves clustering performance by extending the traditional kernel k-means (KKM) clustering algorithm to a multiple setting by combining a group of pre-specified kernels. In this paper, we develop and propose a multiple kernel k-means clustering via latent variables (MKKLV) algorithm, in which base kernels can be adaptively adjusted with respect to each sample. To improve the effectiveness of the kernel-specific and sample-specific characteristics of the data, joint diversity regularization and graph regularization are utilized in the MKKLV algorithm. An efficient three-step iterative algorithm is employed to jointly optimize the kernel-specific and sample-specific coefficients. Experiments validate that our algorithm outperforms state-of-the-art techniques on several different benchmark datasets. [ABSTRACT FROM AUTHOR]

Open data satellite imagery provides valuable data for the planning and decision-making processes related with environmental domains. Specifically, agriculture uses remote sensing in a wide range of services, ranging from monitoring the health of the crops to forecasting the spread of crop diseases. In particular, this paper focuses on a methodology for the automatic delimitation of cultivable land by means of machine learning algorithms and satellite data. The method uses a partition clustering algorithm called Partitioning Around Medoids and considers the quality of the clusters obtained for each satellite band in order to evaluate which one better identifies cultivable land. The proposed method was tested with vineyards using as input the spectral and thermal bands of the Landsat 8 satellite. The experimental results show the great potential of this method for cultivable land monitoring from remote-sensed multispectral imagery. [ABSTRACT FROM AUTHOR]

Image segmentation has been an important technique in the field of image processing. Fine-level manual annotations are very limited and difficult for a large collection of imbalanced images, where each image contains quite different small objects. However, we find that these imbalanced images have similar decision graphs obtained by a lightweight and simple clustering algorithm Density Peaks (DPeaks). Hence, in this paper, we propose a weakly supervised image segmentation algorithm. It trains a decision curve from decision graphs of a few sample imbalanced images by SVM and Support Vector Regression (SVR), which is effective for identifying the sparse region of an imbalanced image. Besides, RangeTree is applied to accelerate RDP for large images due to the high complexity of DPeaks Clustering. Experiments prove that the proposed algorithm works well on imbalanced image data sets, not only it can recognize main things, but also has the ability to identify some relatively small objects. [ABSTRACT FROM AUTHOR]

Clustering is an essential tool in data mining research and applications. It is the subject of active research in many fields of study, such as computer science, data science, statistics, pattern recognition, artificial intelligence, and machine learning. Several clustering techniques have been proposed and implemented, and most of them successfully find excellent quality or optimal clustering results in the domains mentioned earlier. However, there has been a gradual shift in the choice of clustering methods among domain experts and practitioners alike, which is precipitated by the fact that most traditional clustering algorithms still depend on the number of clusters provided a priori. These conventional clustering algorithms cannot effectively handle real-world data clustering analysis problems where the number of clusters in data objects cannot be easily identified. Also, they cannot effectively manage problems where the optimal number of clusters for a high-dimensional dataset cannot be easily determined. Therefore, there is a need for improved, flexible, and efficient clustering techniques. Recently, a variety of efficient clustering algorithms have been proposed in the literature, and these algorithms produced good results when evaluated on real-world clustering problems. This study presents an up-to-date systematic and comprehensive review of traditional and state-of-the-art clustering techniques for different domains. This survey considers clustering from a more practical perspective. It shows the outstanding role of clustering in various disciplines, such as education, marketing, medicine, biology, and bioinformatics. It also discusses the application of clustering to different fields attracting intensive efforts among the scientific community, such as big data, artificial intelligence, and robotics. This survey paper will be beneficial for both practitioners and researchers. It will serve as a good reference point for researchers and practitioners to design improved and efficient state-of-the-art clustering algorithms. • Provide an up-to-date comprehensive review of the different clustering techniques. • Highlight novel and most recent practical applications areas of clustering. • Provide a convenient research path for new researchers. • Help experts develop new algorithms for emerging challenges in the research area. [ABSTRACT FROM AUTHOR]

Clustering Algorithms have recently attracted considerable attention in machine learning applications due to their high efficiency. However, the existing clustering methods still have some issues that need to be further solved. For example, most existing methods convert one type of feature into another type, which ignores the specific properties of data. In addition, most of them consider entire features, which may lead to complexity in computation and result in sub-optimal performance. To address the above problems, this paper proposes a novel method for clustering categorical and numerical features based on Feature Space Instance Cluster Closeness Metric (FSICCM). In the first stage, given training data, FSICCM exploits the similarity metric for numerical features. In the second stage, we design a novel metric for categorical features. Meanwhile, we design a new learning algorithm to cluster mixed datasets. Extensive experimental results on benchmark datasets show that the proposed FSICCM outperforms several state-of-art clustering methods in terms of accuracy and efficiency. • Cluster analysis of mixed data based on Feature Space Instance-Cluster Closeness Metric •Rival Penalization and Initialization Adaptation Process for Cluster Analysis •The algorithm trains the number of clusters. •The proposed algorithm removes unnecessary clusters. [ABSTRACT FROM AUTHOR]

• The proposed hybrid method performs over-sampling and under-sampling effectively. • The sampling process is done quickly and efficiently by clustering. • The roulette wheel selects the most suitable samples through probabilities. • The proposed method does not need to set any parameter. • The proposed method reaches to IR =1 for all imbalanced data sets. The problem of imbalanced data set classification is prevalent in the studies of machine learning and data mining. In these kinds of data sets, the number of samples in classes is unequal so that one class has a lot more samples (the majority or negative class) than the other (the minority or positive class). The classical classifiers are ineffective in these conditions because they are biased toward the majority class and ignore the minority class, which is more important. Preprocessing the data distribution before training the classifier is one of the most effective methods to resolve this problem. These methods, balance the data distribution by decreasing the majority class size (under-sampling methods) or increasing the minority class size (over-sampling methods) or combining both of them (hybrid methods). In this paper, we propose an effective and simple hybrid approach based on the density concept and clustering, which is called Clustering and Density-Based Hybrid (CDBH). First, the minority class samples are clustered by the well-known k -means algorithm and their densities in each cluster are obtained. Then, the denser minority samples are selected with more likely to generate the new minority samples. To decrease the majority class size, the k -means algorithm is applied again on the majority class samples to cluster them and compute their densities, like the previous stage. Finally, the denser majority samples will have more chance to choose from the training set, and other samples are removed to balance the data samples distribution between classes. In the experiments, the Support Vector Machine (SVM) classifier is used as the classifier, and F -measure and AUC criteria are employed for evaluation. Also, preprocessing methods are compared in terms of the complexity of the classification model and the over-sampling rate. The results of comparing CDBH and other state of the art methods over 44 imbalanced data sets show the superiority of the proposed CDBH method based on the F -measure criterion. [ABSTRACT FROM AUTHOR]

• A comprehensive classification of VANET clustering algorithms: strength and weakness. • Intelligence-based VANET clustering including machine learning and fuzzy logic. • The distinction between vehicle mobility as well as network mobility strategies. • A detail study of VANET multi-hop clustering strategies. Clustering is an important concept in vehicular ad hoc network (VANET) where several vehicles join to form a group based on common features. Mobility-based clustering strategies are the most common in VANET clustering; however, machine learning and fuzzy logic algorithms are also the basis of many VANET clustering algorithms. Some VANET clustering algorithms integrate machine learning and fuzzy logic algorithms to make the cluster more stable and efficient. Network mobility (NEMO) and multi-hop-based strategies are also used for VANET clustering. Mobility and some other clustering strategies are presented in the existing literature reviews; however, extensive study of intelligence-based, mobility-based, and multi-hop-based strategies still missing in the VANET clustering reviews. In this paper, we presented a classification of intelligence-based clustering algorithms, mobility-based algorithms, and multi-hop-based algorithms with an analysis on the mobility metrics, evaluation criteria, challenges, and future directions of machine learning, fuzzy logic, mobility, NEMO, and multi-hop clustering algorithms. Image, graphical abstract [ABSTRACT FROM AUTHOR]

With the complexity of the 5 G network environment and the diversified requirements for video transmission, multimedia content transmission in different channel environments is a direction worth studying in order to achieve good performance of digital media content transmission systems in an open network environment. Based on the code stream structure characteristics of screen content in HEVC, this paper has proposed a joint source channel coding (JSCC) scheme to study the transmission of compressed video in wireless channels. Combined with the clustering algorithm from the field of artificial intelligence in the environment of wireless channel classification problem, the parameters of the wireless channel and the influencing factors can be used to reduce the noise and then use the FCM (Fuzzy C-Means) clustering algorithm to classify the channel environment. According to the channel status, optimize the current remaining resources and implement channel coding through LDPC codes. By analyzing and minimizing the end-to-end distortion model, the adaptive bit rate allocation further guarantees the quality of the reconstructed video. [ABSTRACT FROM AUTHOR]