基于随机森林和遗传算法的Ceph参数自动调优

Ceph系统性能受Ceph配置参数的显著影响,在Ceph集群的配置优化中,配置参数种类繁多、含义复杂,导致难以实现快速准确寻优。针对以上问题,提出一种基于随机森林（RF）和遗传算法（GA）的参数调优方法,用于自动调整Ceph参数配置以优化Ceph系统性能。该方法使用RF算法为Ceph系统构建性能预测模型,并将预测模型的输出作为GA的输入,通过GA对参数配置方案进行自动迭代优化。仿真结果表明,调优后的参数配置较默认的参数配置相比,使Ceph文件系统的读写性能提高了约1.4倍,并且寻优耗时远低于黑盒参数调优方法。     

改进随机森林及其在人体姿态识别中的应用

针对随机森林算法静态性、容易陷入局部最优等问题,提出了一种蜜蜂交配优化的随机森林算法,并将该算法应用于基于加速度传感器的人体姿态识别。设计了一套以三轴加速度传感器MMA7260与无线通信模块CC2430相结合的数据采集系统,采集了五种日常行为和一种异常行为;从加速度值中提取了近斜率、前后差、均值、均方根和信号幅值面积5类特征矢量;采用改进的随机森林算法训练行为模型和进行分类识别。实验结果表明：该算法能有效地识别六种行为,具有较高的分类预测准确率和行为识别率,且具有较强的稳定性、鲁棒性、全局寻优和抗噪声能力。     

Recognizing signal trends on-line by a fuzzy-logic-based methodology optimized via genetic algorithms

The present work addresses the problem of on-line signal trend identification within a fuzzy logic-based methodology previously proposed in the literature. A modification in the application of the methodology is investigated which entails the use of singletons instead of triangular fuzzy numbers for the characterization of the truth values of the six parameters describing the dynamic trend of the evolving process. Further, calibration of the model is performed by a genetic algorithm procedure. In an example of application of the method, this procedure is also exploited for feature selection, i.e. for choosing which of the measured plant signals are relevant for the transient identification.     

基于遗传算法选优的集成手段与时序卷积网络的涡扇发动机剩余寿命预测

涡扇发动机作为航空航天领域的核心设备之一,其健康状况决定了航空器能否稳定可靠地运行。而对涡扇发动机的剩余寿命（RUL）进行判断,是设备监测与维护的重要一环。针对涡扇发动机监测过程中存在的工况复杂、监测数据多样、时间跨度长等特点,提出了一种遗传算法优选时序卷积网络（TCN）基模型的集成方法（GASEN-TCN）的涡扇发动机剩余寿命预测模型。首先,利用TCN捕获长跨度下的数据内在关系,从而对RUL作出预测;然后,应用GASEN集成多个独立的TCN,以增强模型的泛化性能;最后,在通用的商用模块化航空推进系统模拟模型（C-MAPSS）数据集上,对所提模型与当下流行的机器学习方法和其他的深度神经网络进行了比较。实验结果表明,在多种不同的运行模式和故障条件下,与流行的双向长短期记忆（Bi-LSTM）网络相比,所提模型都有着更高的预测准确率与更低的预测误差。以FD001数据集为例,在该数据集上所提模型的均方根误差（RMSE）相较Bi-LSTM低17.08%,相对准确率（Accuracy）相较Bi-LSTM高12.16%。所提模型在设备的智能检修与维护方面有着较好的应用前景。     

基于遗传算法选优的集成手段与时序卷积网络的涡扇发动机剩余寿命预测

涡扇发动机作为航空航天领域的核心设备之一，其健康状况决定了航空器能否稳定可靠地运行。而对涡扇发动机的剩余寿命（RUL）进行判断，是设备监测与维护的重要一环。针对涡扇发动机监测过程中存在的工况复杂、监测数据多样、时间跨度长等特点，提出了一种遗传算法优选时序卷积网络（TCN）基模型的集成方法（GASEN-TCN）的涡扇发动机剩余寿命预测模型。首先，利用TCN捕获长跨度下的数据内在关系，从而对RUL作出预测；然后，应用GASEN集成多个独立的TCN，以增强模型的泛化性能；最后，在通用的商用模块化航空推进系统模拟模型（C-MAPSS）数据集上，对所提模型与当下流行的机器学习方法和其他的深度神经网络进行了比较。实验结果表明，在多种不同的运行模式和故障条件下，与流行的双向长短期记忆（Bi-LSTM）网络相比，所提模型都有着更高的预测准确率与更低的预测误差。以FD001数据集为例，在该数据集上所提模型的均方根误差（RMSE）相较Bi-LSTM低17.08%，相对准确率（Accuracy）相较Bi-LSTM高12.16%。所提模型在设备的智能检修与维护方面有着较好的应用前景。     

Recognizing human actions by learning and matching shape-motion prototype trees

A shape-motion prototype-based approach is introduced for action recognition. The approach represents an action as a sequence of prototypes for efficient and flexible action matching in long video sequences. During training, an action prototype tree is learned in a joint shape and motion space via hierarchical K-means clustering and each training sequence is represented as a labeled prototype sequence; then a look-up table of prototype-to-prototype distances is generated. During testing, based on a joint probability model of the actor location and action prototype, the actor is tracked while a frame-to-prototype correspondence is established by maximizing the joint probability, which is efficiently performed by searching the learned prototype tree; then actions are recognized using dynamic prototype sequence matching. Distance measures used for sequence matching are rapidly obtained by look-up table indexing, which is an order of magnitude faster than brute-force computation of frame-to-frame distances. Our approach enables robust action matching in challenging situations (such as moving cameras, dynamic backgrounds) and allows automatic alignment of action sequences. Experimental results demonstrate that our approach achieves recognition rates of 92.86 percent on a large gesture data set (with dynamic backgrounds), 100 percent on the Weizmann action data set, 95.77 percent on the KTH action data set, 88 percent on the UCF sports data set, and 87.27 percent on the CMU action data set.     

Multi-scale detection and interpretation of spatio-temporal anomalies of human activities represented by time-series

Spatio-temporal patterns of human activities can be affected by events, such as extreme weather. Events cause anomalies that could be expressed by abnormal activity patterns deviating from the inherent ones. The detection of spatio-temporal anomalies thus helps to understand the implicit influencing mechanism with which the external factors affect human activities. Existing methods of spatio-temporal anomaly detection usually treat the temporal information as attributes of spatial units, which is an over-simplification as it ignores complex temporal patterns (e.g., periodic components of time-series). Moreover, as the spatio-temporal resolutions affect expressed characteristics of anomalies, the sensitivity of anomalies to scale is also worth investigating. This study intends to detect and interpret the spatio-temporal anomalies of human activities from a multi-scale perspective. Being different from the single-scale consideration and independent consideration of multiple scales, this research investigates how the anomalies' characteristics change at multiple scales by anomaly matching. The criteria of anomaly matching are the overlapping degree of spatio-temporal influence ranges of anomalies. It helps to specify the events that caused the expressed anomalies. Besides, we introduce the time-series decomposition methods to decompose complex temporal patterns, highlighting the abnormal changes in activity patterns. The study is validated using a multi-temporal-scale simulation experiment, and a multi-spatial-scale experiment based on taxi data in Beijing. Results show that the multi-scale method can detect various anomalies. Moreover, obtained multi-scale characteristics of anomalies are easy to compare with external data, and thus benefit anomaly interpretation (validated by two sample anomalies). This study highlights the significance of scales in anomaly detection of human activities and provides references for related works.     

Multi-objective optimization by genetic algorithm of structural systems subject to random vibrations

This paper deals with a multi-objective optimization criterion for linear viscous-elastic device utilised for decreasing vibrations induced in mechanical and structural systems by random loads. The proposed criterion for the optimum design is the minimization of a vector objective function. The multi-objective optimization is carried out by means of a stochastic approach. The design variables are the device frequency and the damping ratio. As cases of study, two different problems are analysed: the base isolation of a rigid mass and the tuned mass damper positioned on a multi degree of freedom structural system subject to a base acceleration. The non-dominated sorting genetic algorithm in its second version (NSGA-II) is adopted to obtain the Pareto sets and the corresponding optima for different characterizations of the system and input.     

Toward a definition and understanding of correlation for variables constrained by random relations

Random relations are random sets defined on a two-dimensional space (or higher). After defining the correlation for two variables constrained by a random relation as an interval, the effect of imprecision was studied by using a multi-valued mapping, whose domain is a space of joint random variables. This perspective led to the notions of consistent and non-consistent marginals, which parallel those of epistemic independence, and unknown interaction and epistemic independence for random sets, respectively. The calculation of the correlation bounds entails solving two optimisation problems that are NP-hard. When the entire random relation is available, it is shown that the hypothesis of non-consistent marginals leads to correlation bounds that are much larger (four orders of magnitude in some cases) than those obtained under the hypothesis of consistent marginals; this hierarchy parallels the hierarchy between probability bounds for unknown interaction and strong independence, respectively. Solutions of the optimisation problems were found at the extremes of their feasible intervals in 80–100% of the cases when non-consistent marginals were assumed, but this range became 75–84% when consistent marginals were assumed. When only the marginals are available, there is a complete loss of knowledge in the correlation, and the correlation interval is nearly vacuous or vacuous (i.e. [ ? 1,1]) even if the measurements are sufficiently accurate in which their narrowed intervals do not overlap. Solutions to the optimisation problems were found at the extremes of their feasible intervals 50% or less of the times.     

Integrated optimization of inventory-distribution systems by random local search and a genetic algorithm

A new model and its solution procedure for the commodity distribution system consisting of distribution centers and consumer points are discussed. Demand is assumed to be a random variable that obeys a known, stationary probability distribution. An integrated optimization model is built where both the order-up-to-R policy, which is one of the typical inventory policies for periodic review models, and the transportation problem are considered simultaneously. The assignment of consumer points to distribution centers is not fixed. The problem is to determine the target inventory and the transportation quantity in order to minimize the expectation of the sum of inventory related costs and transportation costs. Simulation and linear programming are used to calculate the expected costs, and a random local search method is developed in order to determine the optimum target inventory. A genetic algorithm is also tested and compared with the proposed random local search method. The model and effectiveness of the proposed solution procedure are clarified by computational experiments.     

RFRCDB-siRNA: improved design of siRNAs by random forest regression model coupled with database searching

Although the observations concerning the factors which influence the siRNA efficacy give clues to the mechanism of RNAi, the quantitative prediction of the siRNA efficacy is still a challenge task. In this paper, we introduced a novel non-linear regression method: random forest regression (RFR), to quantitatively estimate siRNAs efficacy values. Compared with an alternative machine learning regression algorithm, support vector machine regression (SVR) and four other score-based algorithms [A. Reynolds, D. Leake, Q. Boese, S. Scaringe, W.S. Marshall, A. Khvorova, Rational siRNA design for RNA interference, Nat. Biotechnol. 22 (2004) 326-330; K. Ui-Tei, Y. Naito, F. Takahashi, T. Haraguchi, H. Ohki-Hamazaki, A. Juni, R. Ueda, K. Saigo, Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference, Nucleic Acids Res. 32 (2004) 936-948; A.C. Hsieh, R. Bo, J. Manola, F. Vazquez, O. Bare, A. Khvorova, S. Scaringe, W.R. Sellers, A library of siRNA duplexes targeting the phosphoinositide 3-kinase pathway: determinants of gene silencing for use in cell-based screens, Nucleic Acids Res. 32 (2004) 893-901; M. Amarzguioui, H. Prydz, An algorithm for selection of functional siRNA sequences, Biochem. Biophys. Res. Commun. 316 (2004) 1050-1058) our RFR model achieved the best performance of all. A web-server, RFRCDB-siRNA (http://www.bioinf.seu.edu.cn/siRNA/index.htm), has been developed. RFRCDB-siRNA consists of two modules: a siRNA-centric database and a RFR prediction system. RFRCDB-siRNA works as follows: (1) Instead of directly predicting the gene silencing activity of siRNAs, the service takes these siRNAs as queries to search against the siRNA-centric database. The matched sequences with the exceeding the user defined functionality value threshold are kept. (2) The mismatched sequences are then processed into the RFR prediction system for further analysis.     

Recognizing human gestures in videos by modeling the mutual context of body position and hands movement

In our days, due the evolution of high-speed computers, the old Human–Computer Interface (HCI) legacies based on mouse and keyboard are slowly becoming obsolete and cannot be accurate enough and respond in a timely manner to the flow of information today. This is why new ways of communicating with the computer have to be researched, the most natural one being the use of gestures. In this paper, a two-level architecture for recognizing human gestures from video frames is proposed. The architecture makes use of several feed-forward neural networks to compute the gestures based on the Haar-like features of body, hand and finger as well as a stochastic-free context grammar that is employed to comprise the mutual context between body pose and hand movement. Trained and tested on 10 gestures (Swipe Right, Swipe Left, Swipe Up, Swipe Down, Horizontal Wave, Vertical Wave, Circle, Point, Palm Up and Fist) the over 94 % accuracy of the system surpasses the current state of the art and compared with a system with no mutual context between body position and hand movement our proposed architecture shows an increase in accuracy with up to 7 %.     

Support Vector machine and duration-aware conditional random field for identification of spatio-temporal activity patterns by combined indoor positioning and heart rate sensors

Tracking the spatio-temporal activity is highly relevant for domains like security, health, and quality management. Since animal welfare became a topic in politics and legislation locomotion patterns of livestock have received increasing interest. In contrast to the monitoring of pedestrians cattle activity tracking poses special challenges to both sensors and data analysis. Interesting states are not directly observable by a single sensor. In addition, sensors must be accepted by cattle and need to be robust enough to cope with a rough environment. In this article, we introduce the novel combination of heart rate and positioning sensors. Attached to neck and chest they are less interfering than accelerometers at the ankles. Exploiting the potential of such combined sensor system that records locomotion and non-spatial information from the heart rate sensor however is challenging. We introduce a novel two level method for the activity tracking focused on the duration and sequence of activity states. We combine Support Vector Machine (SVM) with Conditional Random Field (CRF) and extend Conditional Random fields by an explicit representation of duration. The SVM characterizes local activity states, whereas the CRF addresses sequences of local states to sequences incorporating spatial and non-spatial contextual knowledge. This combination provides a reliable and comprehensive identification of defined activity patterns, as well as their chronology and durations, suitable for the integration in an activity data base. This data base is used to extract physiological parameters and promises insights into internal states such as fitness, well-being and stress. Interestingly we were able to demonstrate a significant correlation between resting pulse rate and the day of pregnancy.     

Feature selection by genetic algorithms in object-based classification of IKONOS imagery for forest mapping in Flanders, Belgium

Obtaining detailed information about the amount of forest cover is an important issue for governmental policy and forest management. This paper presents a new approach to update the Flemish Forest Map using IKONOS imagery. The proposed method is a three-step object-oriented classification routine that involves the integration of 1) image segmentation, 2) feature selection by Genetic Algorithms (GAs) and 3) joint Neural Network (NN) based object-classification. The added value of feature selection and neural network combination is investigated. Results show that, with GA-feature selection, the mean classification accuracy (in terms of Kappa Index of Agreement) is significantly higher (p < 0.01) than without feature selection. On average, the summed output of 50 networks provided a significantly higher (p < 0.01) classification accuracy than the mean output of 50 individual networks. Finally, the proposed classification routine yields a significantly higher (p < 0.01) classification accuracy as compared with a strategy without feature selection and joint network output. In addition, the proposed method showed its potential when few training data were available.     

Markerless human motion capture by Markov random field and dynamic graph cuts with color constraints

Currently, many vision-based motion capture systems require passive markers attached to key locations on the human body. However, such systems are intrusive with limited application. The algorithm that we use for human motion capture in this paper is based on Markov random field (MRF) and dynamic graph cuts. It takes full account of the impact of 3D reconstruction error and integrates human motion capture and 3D reconstruction into MRF-MAP framework. For more accurate and robust performance, we extend our algorithm by incorporating color constraints into the pose estimation process. The advantages of incorporating color constraints are demonstrated by experimental results on several video sequences. Supported by the National Basic Research Program of China (Grant No. 2006CB303105)