多层传感器故障数据的挖掘模型仿真

多层传感器的故障准确定位对保证各自应用安全至关重要。多层传感器不同于传统的传感器网络,其不同层次的传感器故障的特征差异较大,不同层次传感器之间存在故障特征"断层"问题。传统的基于流数据异常特征识别的多层传感器故障数据的挖掘模型需要明确层次网络故障之间的关联特征,若传感器层次之间的故障特征关联性不强,故障挖掘的阀值就无法固定,产生故障特征无法定位问题,导致误警率较高。提出了一种基于贝叶斯信念网络的多层传感器故障数据的挖掘模型,针对多层传感器故障数据属性多样性的问题,分析了贝叶斯信念网络的结构,搜索一个最匹配待分类故障数据样本的贝叶斯信念网络,通过评估函数评估各个可能的网络结构与样本多层传感器故障数据间的契合度,采集一个最佳样本多层传感器故障数据解,通过"压缩侯选"的贝叶斯信念网络算法,计算样本多层传感器故障数据间的依赖关系,集中扫描最可能是待挖掘数据的变量集,实现故障数据的挖掘。实验结果表明,利用所提模型能够有效提高多层传感器故障数据的挖掘的准确性。

基于小数据冲突检测的坏点数据挖掘模型仿真

传统基于离群度的坏点检测方法,无法解决小数据冲突过程中存在的震荡波动以及数据特征不明显的问题,获取的坏点结果存在较大的偏差。提出了一种基于小数据冲突检测的坏点数据挖据模型,通过小区域异常因子LOD描述小数据冲突数据集中不同数据对象的局部异常程度,采用小数据冲突数据的邻域查询优化算法,获取初步坏点数据集,通过运算小区域异常因子的方法,在邻域搜索优化后获取的对象邻域中,基于两个对象间的加权考斯基距离,采用去一划分信息熵增量获取小数据冲突对象的权值,运算初步坏点数据集中小数据冲突对象的损坏程度,获取小数据冲突中的坏点数据。实验结果说明,所提方法在挖据小数据冲突中的坏点数据过程中,在繁琐度和差异性方面较传统模型都具有较高的优越性。

Gaussian mixture models for clustering and classifying traffic flow in real-time for traffic operation and management

Based on Gaussian mixture models（GMM）, speed, flow and occupancy are used together in the cluster analysis of traffic flow data. Compared with other clustering and sorting techniques, as a structural model, the GMM is suitable for various kinds of traffic flow parameters. Gap statistics and domain knowledge of traffic flow are used to determine a proper number of clusters. The expectation-maximization （E-M） algorithm is used to estimate parameters of the GMM model. The clustered traffic flow pattems are then analyzed statistically and utilized for designing maximum likelihood classifiers for grouping real-time traffic flow data when new observations become available. Clustering analysis and pattern recognition can also be used to cluster and classify dynamic traffic flow patterns for freeway on-ramp and off-ramp weaving sections as well as for other facilities or things involving the concept of level of service, such as airports, parking lots, intersections, interrupted-flow pedestrian facilities, etc.

The multiscale modeling and data mining of high-temperature dielectrics of SiO_2/SiO_2 composites

The high temperature dielectrics of Quartz fiber-reinforced silicon dioxide ceramic （Si02/SiO2 ） composites were studied both theoretically and experimentally. A multi-scale theoretical model was developed based on the theory of dielectrics. It was realized to predict dielectric properties at higher temperature （ 〉 1200 ℃） by experimental data mining for correlative coefficients in model. The results show that the dielectrics of SiO2/SiO2, which were calculated with the theoretical model, were in agreement with experimental measured value.

Research on the Distributed Data Mining Cloud Framework Oriented Internet of Things

The rapid development of Internet of Things imposes new requirements on the data mining system, due to the weak capability of traditional distributed networking data mining. To meet the needs of the Internet of Things, this paper proposes a novel distributed data-mining model to realize the seamless access between cloud computing and distributed data mining. The model is based on the cloud computing architecture, which belongs to the type of incredible nodes.

Optimization of support vector machine power load forecasting model based on data mining and Lyapunov exponents

According to the chaotic and non-linear characters of power load data,the time series matrix is established with the theory of phase-space reconstruction,and then Lyapunov exponents with chaotic time series are computed to determine the time delay and the embedding dimension.Due to different features of the data,data mining algorithm is conducted to classify the data into different groups.Redundant information is eliminated by the advantage of data mining technology,and the historical loads that have highly similar features with the forecasting day are searched by the system.As a result,the training data can be decreased and the computing speed can also be improved when constructing support vector machine(SVM) model.Then,SVM algorithm is used to predict power load with parameters that get in pretreatment.In order to prove the effectiveness of the new model,the calculation with data mining SVM algorithm is compared with that of single SVM and back propagation network.It can be seen that the new DSVM algorithm effectively improves the forecast accuracy by 0.75%,1.10% and 1.73% compared with SVM for two random dimensions of 11-dimension,14-dimension and BP network,respectively.This indicates that the DSVM gains perfect improvement effect in the short-term power load forecasting.

Bankruptcy Prediction for Chinese Firms： Comparing Data Mining Tools With Logit Analysis

China＇s capital market is different from that of the US in economic, political, and socio-cultural ways. China＇s dynamic and fast growing economy for the past decade entails some structural changes and weaknesses and as a consequence, there are some business failures. We propose bankruptcy prediction models using Chinese firm data via several data mining tools and traditional logit analysis. We used Chinese firm data one year prior to bankruptcy and our results suggest that the financial variables developed by Altman （1968） and Ohlson （1980） perform reasonably well in determining business failures of Chinese firms, but the overall prediction rate is low compared with those of the US or other countries＇ studies. The reasons for this low prediction rate may be structural weaknesses resulting from China＇s fast growth and immature capital market.

Comparsion analysis of data mining models applied to clinical research in Traditional Chinese Medicine

OBJECTIVE: To help researchers selecting appropriate data mining models to provide better evidence for the clinical practice of Traditional Chinese Medicine(TCM) diagnosis and therapy.METHODS: Clinical issues based on data mining models were comprehensively summarized from four significant elements of the clinical studies:symptoms, symptom patterns, herbs, and efficacy.Existing problems were further generalized to determine the relevant factors of the performance of data mining models, e.g. data type, samples, parameters, variable labels. Combining these relevant factors, the TCM clinical data features were compared with regards to statistical characters and informatics properties. Data models were compared simultaneously from the view of applied conditions and suitable scopes.RESULTS: The main application problems were the inconsistent data type and the small samples for the used data mining models, which caused the inappropriate results, even the mistake results. These features, i.e. advantages, disadvantages, satisfied data types, tasks of data mining, and the TCM issues, were summarized and compared.CONCLUSION: By aiming at the special features of different data mining models, the clinical doctors could select the suitable data mining models to resolve the TCM problem.

Unconventional dark matter models: a brief review

Although weakly interacting massive particle （WIMP） scenario is very well motivated, it is not guaran- teed to be the truth. It is important to keep mind open and consider other well-motivated scenarios. In this paper, we briefly review several possible non-WIMP dark matter （DM） candidates. First, we discuss details on asymmetric DM models, in which the baryon asymmetry in standard model sector is related to the asymmetry in DM sector. We discuss how DM relic abundance is determined in such models. Also we cover the possible interesting ex- perimental signatures induced by its asymmetric nature. Then we consider ultralight DM candidates, i.e., axion and dark photon. In such scenarios, DM should be treated as a coherently oscillating background, instead of each individual particle. Searching strategies for such DM candidates is very different than those in conventional DM models. We discuss several interesting experiments looking for these ultralight particles. We also cover interesting subtleties encountered in those experiments.