Search

Most of the scientific research engines utilize the same way to rank scientific research papers (SRP), this way is highly depend on the citation network. The retrieved results of these engines for any keyword contain old papers with the highest citation value. In this paper, we proposed a new and easy to implement ranking algorithm for scientific research papers known as Paper Time Ranking Algorithm (PTRA). PTRA is a new ranking algorithm that depends on three factors to rank its results; paper age, citation index and publication venue. To construct and prove our ranking algorithm we created a web crawler that crawl different scientific search papers databases in the world to collect the information that PTRA needs. Unfortunately, some of this information was missing, such as, the impact factor of the journals. To collect this information, we created another crawler that search the Internet for these journals impact factors. However, to prove our ranking algorithm results, we have compared PTRA with Google scholar ranking algorithm. We used Google scholar, since it has more than 50 million papers and its own system can gather and collect papers in a faster way. Our comparison results show that, citation index has the highest impact on Google scholars ranking algorithm results. Unlike PTRA that gains the paper age a higher impact on the ranked results. However, PTRA depends on the citation index and publication venue to rank the results but with less impact than the paper age.

The detection of elements in top part of research paper is very important, because these elements are often used as the search items by user. This paper provides a mixed method for auto detection of top part from research paper. The papers feature of keyword, layout and content similarity are mixed to accurately find the area of top part and recognize the elements in top part. Experiments show the advantage of our method over existing methods, and future work is also described in the paper.

According to the problem that the flexibility and efficiency are not high of the current test paper generating system, applying data mining to figure out the inter and intra connection among all the knowledge points of CET-4, mining association rules, and picking out as more potential weaknesses of knowledge points as it can, then selecting items corresponding to the weaknesses and generating a user-level specific suitable test paper. Experiments show that the new test paper generating algorithm (TPGA) based on association rules response intelligently and quickly to different user levels, and provide an essential improvement in test paper generating process for users own good to enhance their exercise results.

Most of the product design on the market is variant design or adaptive design, which need to reuse existing product design knowledge. A key aspect of reusing existing CAD model is correctly define and understand the design intents behind of existing CAD model, and this paper introduces a CAD model annotation system based on design intent. Design intents contained all design information of entire life cycle from modeling, analysis to manufacturing are marked onto the CAD model using PMI module in UG to improve the readability of the CAD model. Second, given the problems such as management difficulties, no filter and retrieval functions, this paper proposes an annotation manager system based on UG redevelopment by filtration, retrieval, grouping and other functions to reduce clutter on the 3D annotations and be convenient for users to view needed all kinds of annotations. Finally, design information is represented both internally within the 3D model and externally on a XML file.

An increasing number of companies establish energy management systems for continuous improvement in their energy efficiency and for this intensify monitoring their current energy consumption. These data can be used to gain further information about the production and to find potentials to increase its energy efficiency. In the procurement process of machinery and equipment or in the planning phase of production systems and building services, information about energy demand is rarely available, though it would be valuable for an early inclusion of energy efficiency in these processes. Therefore this paper discusses different forecast methods for energy consumption of machinery and evaluates in particular their universal applicability, effort and accuracy by analyzing them through the example of a packaging machine. In addition this paper proposes a further usage of energy-related data of machinery, which can be automatically acquired by monitoring systems for prognosticating their energy consumption as well as a possible distribution approach of this information. Therefore an own forecast method is presented, which shall process the energy-related data combined with information about dominant parameters of the product, the usage of the machine and the environmental conditions. For the distribution concept it was taken into account that the generated and shared information has to be abstracted in a way that no critical secrets of the company are revealed.

The paper deals with selected approaches which unite several correlation analysis principles. Field data very often has various forms and contents. The comparison of different approaches will help to determine more precisely which correlation analysis is better for assessing input and output data. In this paper we introduce several correlation principles which can help to select the most suitable correlation approach. We present a traditional correlation analysis and compare it with Pearson and Spearman correlation coefficients. The value of our article lies in comparing several different approaches of the correlation analysis in which the oil field data from diesel combustion engine are used

Creation of the intelligent system for highly structured domain is possible with determining the whole system of concepts included in the ontology of this area. Presentation by means of formal ontology model provides a uniquely interpreted terms. Based on the model of ontology the intelligent system will integrate their knowledge and information components of all sections of the data area. The paper considers the unification of nanomaterials subdomain with other subdomains that are need to be used for intelligent system for nanotechnology. The paper also proposes mathematical ontology model of one of the component - that is ontology model of foams and emulsions class.

In this paper, we present a new knowledge push technology for complex mechatronic products design based on ontology and variable precision rough set (VPRS). Ontology can explicitly represent knowledge semantics and let designers exchange knowledge about design and the product development process; while using the knowledge reduction method based on the VPRS method, the design knowledge repository is simplified and design rules also be distilled from the reduct design knowledge repository. On the basis of the above approach, multiple designers can efficiently share design knowledge and can obtain appropriate design knowledge during all design processes. Finally, a case is employed to validate the proposed method of this paper.

The following paper presents an IT tool for aiding assessment of a safety culture level in an organization in a form of an application, providing fast and easy way to analyze data obtained from questionnaires. Traditional data analysis often turns out to be time-consuming and labor-intensive. To make the process faster, more detailed and to find all the possible links among all the available data there is a need for an IT tool. The presented application has been developed with publicly available and completely free tools and technologies. It is a complete piece of software, fully flexible which can be modified or equipped with new functions (modules). The presented solution eliminates the time consuming process of “copying” data from paper questionnaires, and allows to get information about the safety culture level right after adding a questionnaire to a questionnaire database. The presented tool can function as a mobile questionnaire which can be used in any place with Internet access and to fill it in only a browser is required.

An intrusion is a series of relevant actions that occur to a victim in some sequence through the Internet. In this paper, a serial episode mining is first applied to find all possible sophisticated Internet attacks, and then an episode pruning skill is applied to cut unnecessary ones to reduce administrator’s further effort. Input data, log files from a honeypot system, is regarded as a sequence of events, where each event has an associated time of occurrence. The method proposed in this paper can be used to detect abnormal Internet episodes including unknown attacks. Some experiments had been conducted to show the effectiveness of the proposed method.

In this paper, several data mining techniques were discussed and analyzed in order to achieve the objective of human daily activities recognition based on a continuous sensing data set. The data mining techniques of decision tree, Naïve Bayes and Neural Network were successfully applied to the data set. The paper also proposed an idea of combining the Neural Network with the Decision Tree, the result shows that it works much better than the typical Neural Network and the typical Decision Tree model.

This paper puts forward an ensemble model based on uncertainty simulation and multiple model combination for short-term traffic flow prediction. Firstly, the initial condition of traffic flow be simulated by using a normal distribution, which makes a better description of the real traffic flow situation. Secondly, an ensemble prediction model with different types of prediction methods has been presented in order to catch any possible change of trend in the traffic flow. In the ensemble model, In order to improve the prediction accuracy, this paper put forwards an equivalence test and dispersion adaptability test for choosing the most effective methods in the ensemble system. Finally, a case study be given to show the performance of the ensemble model. Predication result shows that this model has a good performance with freeway traffic flow, It is capable of providing more detail about the traffic volume, it provide mean value and standard deviation value of traffic volume in the next period, which is most important information for traffic managers and travelers.

Wind power prediction is a key problem in optimizing power dispatching. This paper builds a wind power prediction model based on wavelet neural network which substitutes wavelet basis function for the transfer function of hidden layer. A missing data interpolation strategy is also given to improve the applicability of the model. With the wind farm data from southeast coast, the model works and the wind power in the next 30 hours is predicted. In the sense of the mean square errors this paper compared the prediction results of the model and BP neural network model, the results shows the models have a better accuracy.

This paper presents a lightweight index does not suspend services online update program, and demonstrate the performance of the index update program from the theoretical analysis and experimental data in two ways. A new method of MapReduce existing index methodology based on this design and further discussion are done in the paper, in the index MapReduce and Hadoop MapReduce feasibility aspects, the design flaws through experimentation.

Dynamic optimization scheduling of the gas in iron and steel enterprises has great significance to reduce gas emission and the short-term forecast is the premise to realize the energy dynamic scheduling. Based on the characteristics that the influencing factors of blast furnace gas amount are complex and difficult to collect, a grey radial basis function (RBF) neural network forecast model is proposed to predict the gas amount for blast furnace in this paper. Combining grey theory, which is used to preprocess the historical data and obtain abundant information, with RBF neural network makes the effective trend forecast in the next 30 minutes come true. The model proposed in this paper is proved to be more accurate according to control experiments against the grey BP neural network.

The operation of power system generates a huge number of data. Manual screening of bad data is almost impossible to achieve. Bad data detection of the power system is the essential prerequisite for power system state estimation. So combined with the characteristics of bad data of the power system, this paper puts forwards an improved K-Means algorithm based on PSO. The traditional K-Means algorithm has disadvantage that it has low efficiency to determine the cluster’s center, but PSO algorithm has the ability to determine the cluster’s center quickly and accurately, this paper combines them together and take advantages of each of the two. The improved K-Means algorithm is based on PSO and verified the effectiveness of the algorithm. The results satisfy the requirements of screening bad data and reduce the false alarms and the probability of undetected and improve the recognition efficiency.

Rapid prototyping technology can generate parts directly from the three-dimensional CAD entity data, hierarchical algorithm is a key link in rapid prototyping manufacture. In this paper, mainly used a hierarchical approach in RP equipment,for the study aimed at three commonly hierarchical algorithm (hierarchical algorithm based on STL model, direct sliced of CAD models, Smart Slice based on STL model) by Perspective, trying to provide a reference for the stratification algorithm of RP system by comparison of the three tiered approach,This paper argues that the Smart slicing algorithm based on STL model is more accurate stratification algorithm

The paper digs out useful conclusions for businesses from the mass of product reviews. To conduct commodity pricing, merchandise improvements provide a basis so as to promote sales performance. The paper discusses the product reviews of emotional connotation. It summed up the current situation development of emotional analysis on product reviews. It analyzes the significant commercial value of product reviews. Product reviews explores the relationship between emotion and sales performance analysis. We can research the technical implementation process of reviews.

Current prediction models for network traffic cannot accurately depict the multi-properties of the Internet traffic. This paper proposes a wavelet-based hybrid model prediction method for network traffic called CLWT model and proposes a prediction method for traffic based on this model. The traffic time series can be rapidly decomposed respectively into approximate time series and detail time series with LF and HF response. The approximate time series predicts by making use of Least Squares Support Vector Machine and proceeds error calibration by using Generalized Recurrent Nerve Network. The detail time series predict it by making use of self-adaption chaotic prediction methods after the medium-soft threshold noise reduction. Finally the prediction value of time series is got by making use of promoting wavelet reconstitution. The effectiveness for the prediction methods mentioned in the paper has been validated by simulation experiment. High prediction accuracy is obtained compared with the existing methods.

This paper investigates the short-term wind power forecasting and demonstrates accurate modeling, which utilizes two representative heuristic algorithms (i.e. wavelet neural network (WNN) and Multilayer Perceptron (MLP)), and statistical machine learning techniques (i.e. Support Vector Regression (SVR)). The proposed method generates the performances of different approaches for random time series, characterized with high accuracy and high generalization capability. The employed data is obtained through Sampling equipment in Real Wind Power Plants (Power generation equipment is Dongfang Steam Turbine Co., Ltd. weak wind turbine type--FD77 with German REpower company technology). The main innovation of this paper comes from: (a) problem may encounter in the real application is in consideration such as corrupt, missing value and noisy data. (b) Data lag estimation are provided to investigate the data distribution and obtain the best input variables, respectively. (c) Comparison between MLP neural networks, WNN and SVR with optimized kernel parameters based on Grid-search method are provided to demonstrate the best forecasting approaches. The purpose of this paper is to provide a method with reference value for short-term wind power forecasting.

This paper introduces a methodology for creating geometrically of real terrain based on DEM data, which provided by the international scientific data sharing center.The method is to convert the DEM data in Global Mapper, AutoCAD and other softwares, process batch of data using office software, based on APDL language programming in ANSYS software , so as to realize the real remote sensing data to establish 3-D geological model in finite element software. An example illustrates the capabilities of the proposed methodology in this paper.

In order to ensure the high performance, the high reliability, the high availability and the high service quality of network, fault management has become an urgent problem to be solved. The application of Rule-Based Reasoning (RBR) and Case-Based Reasoning (CBR) is recognized as an economically viable network fault process technology. To meet the demands of network fault management, this paper presents an association analysis and processing strategy for network events based on the combination of RBR and CBR. The network events are firstly passed through pretreatment module, the useful ones are saved for the following processing. Then, the first processing strategy which is more appropriate is select by RBR/CBR selector. Besides, a new CBR processing strategy is proposed by this paper. By the use of the algorithm based on the preliminary screening and the calculating of variance, the processing efficiency of CBR improved. At last, if the alarm event is not been solved, the system will alarm the related people and the solution can be used to rich the rule base and the case base. Based on these steps, experimental results are given to verify the proposed strategy.

This paper introduces the concept of database and data mining, combined with management system of quality assessment system and method of data mining technology. In this paper, applying the data mining skill to the field of remote open management system, introduces the development of data mining in China and the necessity and importance of data mining in remote open information management system. This thesis analyzes the main problems in the remote open management system. On the basis of the relevant researches both at home and abroad, it presents the significance of the application of data mining in remote open management system. It analyzes the needs of the system based on data mining and presents a detailed design and implication of such a system.

this paper uses a new genetic algorithm (New Genetic Algorithm, NGA) to implement the automatic group volume function, solving the problem that the paper is not fully considered scores distribution of knowledge points in the test paper (test dimension) in the system using traditional genetic algorithm (Genetic Algorithm, GA) to implement automatic group volume, putting forward and redefining the fitness function to speed up the convergence. Simulation experiments show that the NGA algorithm is not only efficient, but can generate a valid paper, making the paper score can on the multi-dimensional of the test as far as possible to achieve uniform distribution.

This paper investigates the application of systematic cluster, one of the cluster analysis methods, in the IED switch online condition monitoring. So far, problems still exsit in the IED switch online condition monitoring based on the control system of smart substation. In order to figure out a relatively stable online model and make classification from the online information of IED switch, this paper firstly fliters and preprocesses the online condition information of IED switch, and uses data matrix to describe the online condition of IED switch. Based on this, the paper inroduces a special dissimilarity degree formula to transfer the data matrix into dissimilarity degree matrix, and conducts analysis with clustering pedigree chart out of systematic cluster algorithm. Finally, the paper verifies the feasibility of systematic cluster in the IED switch online condition monitoring.

Power load forecasting is one of the important parts in power system planning. This paper focused on the research of load forecasting and proposed an improved Rough set for mid-long term load forecasting. Based on the theory, the most sensitive factor can be found, which is applied to studying the uncertain problems of mid-long term load forecasting in this paper and one method fitting for mid-long term load forecasting are proposed. The validity and effectiveness of the method is tested in a real power system example.

In order to make effective use a large amount of graduate data in colleges and universities that accumulate by teaching management of work, the paper study the data mining for higher vocational graduates database using the data mining technology. Using a variety of data preprocessing methods for the original data, and the paper put forward to mining algorithm based on commonly association rule Apriori algorithm, then according to the actual needs of the design and implementation of association rule mining system, has been beneficial to the employment guidance of college teaching management decision and graduates of the mining results.

Classification is a supervised learning problem used to discriminate data instances in different classes. The solution to this problem is obtained through algorithms (classifiers) that look for patterns of relationships between classes in known cases, using these relationships to classify unknown cases. The performance of the classifiers depends substantially of the data types. In order to give proper treatment to nominal data, this paper shows that the application of previous transformations can substantially improve the performance of classifiers, bringing significant benefits to the result of the whole process of Knowledge Discovery in Databases (KDD). This paper uses three different data sets with nominal data and two well-known classifiers: the Linear Discriminant Analysis (LDA), and the Naïve-Bayes (NB). For data transformation, the paper applies an approach called Geometric Data Analysis (GDA). The GDA techniques compared in this paper are the traditional Principal Component Analysis (PCA) and the underexplored Multiple Correspondence Analysis (MCA). The results confirm the capability of the GDA transformation to improve the classification accuracy and attest the superiority of the MCA in comparison with its precursor, the PCA, when applied to nominal data.

With the development of information technology, the system structure of Oracle database is increasingly complicated. The performance is improved, and the application range is expanded, but there are still many unstable factors. Combined with Oracle system structure, the paper expounds data optimization from many aspects including OFA free structure, rollback segment, operating mechanism, hardware and OS environment, and the paper formulates the corresponding optimization strategies.

With the rapid development of GIS, it is applied to network management gradually. We can get the distributed trend of the topological nodes and the topology relation between each other through GIS. In order to reflect the topology relation when we change the zoom level of the map, it is a critical problem that how to realize a large numbers of topological nodes be aggregated quickly and accurately. In view of this situation, this paper introduced two common aggregation algorithms firstly, and then proposed an aggregation algorithm for aggregating the topological nodes in GIS with the advantages of hierarchical clustering algorithm. The paper also explained the basic idea and realization of the aggregation algorithm. The experiment results show that the algorithm provides an effective solution for aggregating topological nodes.

Association rule mining is an important data mining method; it is the key link of finding frequent itemsets. The process of association rules mining is roughly into two steps: the first step is to find out from all the concentration of all the frequent itemsets; the second step is to obtain the association rules from frequent itemsets. This paper analyzes the collected information of nodes in wireless sensor network and management. The paper presents application of association rule mining technology in the collection and management of wireless sensor network node.

Firstly, the paper proposes the framework of recommendation system based on web usage mining and market mechanism. Then, the paper introduces the system structure. And the paper introduces each Agent of the system including the functions of management Agent, interface Agent, online-recommendation Agent and filter Agent. And the paper gives the operation mechanism, resource allocation mechanism, customization return mechanism and customization bidding strategy

As the development of WoT(Web of Things), a large mount of services emerge. How to make full use of these services is a hot research area nowadays. This paper focuses on making these services more valuable in WoT Smart Home scenario. In this paper, we propose a model called service selection model based on rule and statistics into WoT Smart Home scenario. It is a combination of rule model and statistics model. It not only automatizes the service selection process by adopting statistics algorithm, but also dramatically improves the accuracy of this process by combining rule and statistics. Further more, it saves efforts of the annotation of service rules greatly. We carry out some experiments to verify the efficiency of this model by comparing with rule model and statistics respectively. The results show that the combination model can efficiently and precisely select services.

With the promotion of social information construction and the rapid update and replacement of large capacity storage equipment, the amount of data from every field grows exponentially. Reportedly, the amount of the data accumulated by Shandong Hi-speed Group is very large. These data can satisfy us some daily usefulness, such as query, retrieval, statistics, statements etc. But what is more important is that how can we discover some useful information from the information ocean. This information can be used in real life such as auxiliary decision. This paper is proposed in this historical background. Data mining is a powerful tool for acquiring knowledge from massive data. Some methods of data mining, such as decision tree, support vector machine, Bayesian decision theory, artificial neural network, k-nearest neighbor, association rule mining etc, are commonly used. In this paper, we design a recommendation system of highway ETC card by using the theory of decision tree. The recommendation system can predict whether a car owner is a potential ETC customer or not through the analysis of the vehicle information. Experiments proved that the accuracy rate of the recommendation system is larger than 90%, so it can provide effective information for the extension of China's ETC card.

This paper proposes an evaluation method based on fuzzy analytic hierarchy process (FAHP) for network security risk assessment. The paper analyzes the network security risk factors. By paired comparison of the importance of factors, the interval number fuzzy complementary judgment matrix is established to get the weight vectors of interval numbers of the factors. Then, by comparing the weight vectors of interval numbers, a possibility degree matrix is established, and then the weight coefficient of each factor are obtained. Finally the sorting result of the network security risk factors is presented. Instances show that the method is feasible and effective and it provides a reliable basis for security risk control strategy of network security staff.

This Paper proposes a quick matching method for vehicle plate data based on NoSQL database technology considering the huge amounts of traffic information, the method uses the basic traffic network information database which built based on vehicle license plate auto recognition system and it improves the effectiveness and applicability of huge data-matching. Besides, the paper fits the travel time by the time priority principle, and estimates the average travel time by the mean estimation and median estimation according to sample capacity. The practical application shows that this method can effectively improve the efficiency of the mass data processing, and obtain travel time information quickly.

the accurate data mining problem is studied in this paper. With the increasing of data attributes, degree of complexity of the data storage is also increased, resulting in that in data mining process, the complexity of computation is too high, reducing the convergence of the data mining method, thereby reducing the efficiency of data mining. To this end, this paper presents a data mining method based on association rules algorithm. The data is made simplified processing, to obtain the association rules between data which provides the basis for data mining. According to the association rules between the data, the data in line with the minimum support degree is calculated, to achieve accurate data mining. Experimental results show that the proposed algorithm for data mining, can improve mining efficiency, and achieve the desired results.

At present there are many data mining methods. This paper studies the application of rough set method in data mining, mainly on the application of attribute reduction algorithm based on rough set in the data mining rules extraction stage. Rough set in data mining is often used for reduction of knowledge, and thus for the rule extraction. Attribute reduction is one of the core research contents of rough set theory. In this paper, the traditional attribute reduction algorithm based on rough sets is studied and improved, and for large data sets of data mining, a new attribute reduction algorithm is proposed.

The fault in power system cannot be completely avoided. In this paper, we developed a method to resolve fault localization problems in power system. Though the data acquisition process has been highly automated, the process of assimilating and analyzing data still lags behind. Raw data must be transformed into knowledge in order to help users decide how to respond to the event and implement the necessary actions. A promising technique for substation event analysis using rough set theory is described in this paper. It interprets the data and outputs meaningful and concise information, which improves the performance of a data analysis system and help with the knowledge acquisition process. A substation model was developed to generate various fault scenarios for our case studies to evaluate the performance of the rough set algorithm. The results show that it works well and efficiently with the overwhelming data.

In this paper, the process of the developments and changes of the network intrusion behaviors were analyzed. An improved epidemic spreading model was proposed to study the mechanisms of aggressive behaviors spreading, to predict the future course of an outbreak and to evaluate strategies to control a network epidemic. Based on Artificial Immune Systems, the concepts and formal definitions of immune cells were given. And in this paper, the forecasting algorithm based on Markov chain theory was proposed to improve the precision of network risk forecasting. The data of the Memory cells were analyzed directly and kinds of state-spaces were formed, which can be used to predict the risk of network situation by analyzing the cells status and the classification of optimal state. Experimental results show that the proposed model has the features of real-time processing for network situation awareness.

With the development of network science, the link prediction problem has attracted more and more attention. Among which, link prediction methods based on similarity has been most widely studied. Previous methods depicting similarity of nodes mainly consider their common neighbors. But in this paper, from the view of network environment of nodes, which is to analysis the links around the pair of nodes, derive nodes similarity through that of links, a new way to solve the link prediction problem is provided. This paper establishes a link prediction model based on similarity between links, presents the LE index. Finally, the LE index is tested on five real datasets, and compared with existing similarity-based link prediction methods, the experimental results show that LE index can achieve good prediction accuracy, especially outperforms the other methods in the Yeast network.

With the rapid development of network, the security problem of network becomes an issue which has been paid more and more attentions to. Among so many methods of intrusion prevention, data mining is a very effective one. The FP-growth algorithm is the most widely used algorithm for mining frequent item-sets, which is also an algorithm for mining association rules without candidate set. However, the FP-growth algorithm needs large memory when mining large database,and its running speed is slow. In order to overcome these problems, based on the FP-growth algorithm, this paper proposed an optimized algorithm. This paper compared the new algorithm with the previous one based on intrusion prevention model for campus network by experiments. Based on Experiments, we can draw the conclusion that, mining association rules by using the improved FP-growth algorithm can effectively detect the users’ behavior pattern, historical pattern and the current model to calculate the similarity of users, and provides the possibility to accurately judge the user behavior.

Neural network has obvious advantages on dealing with the uncertain problems with a huge amount of data. Some equipment’s faults performance of such characteristics that the date is abundant, and failure phenomenon is not explicit and uncertain. Then, it leads to being hard to diagnose the faults through the traditional diagnosis method in a short time. This paper will analysis the data feature and then build a model to deal with the qualitative data attributes in order that the BP network can use it smoothly. Calculation result shows that using this method, fault diagnosis can be simply and quickly. The paper also provides a new kind of composite way to figure out fault positions for the front-line operators based on experts’experience knowledge but not on measurement signals.

Frequent pattern mining has been an important research direction in association rules. This paper use a methodology by preprocessing the original dataset using fuzzy clustering which can mapped quantitative datasets into linguistic datasets. Then we propose a algorithm based on fuzzy frequent pattern tree for extracting fuzzy frequent itemset from mapped linguistic datasets. Experimental results show that our algorithm is shorter than the F-Apriori on computing time to huge database. For large database, the algorithm presented in this paper is proved to have a good prospect.

According to the analysis of text feature, the document with co-occurrence words expresses very stronger and more accurately topic information. So this paper puts forward a text clustering algorithm of word co-occurrence based on association-rule mining. The method uses the association-rule mining to extract those word co-occurrences of expressing the topic information in the document. According to the co-occurrence words to build the modeling and co-occurrence word similarity measure, then this paper uses the hierarchical clustering algorithm based on word co-occurrence to realize text clustering. Experimental results show the method proposed in this paper improves the efficiency and accuracy of text clustering compared with other algorithms.

Through the software defect prediction can effectively guide the rational distribution of software system development resources, so as to improve the quality of software and software reliability. In order to fully utilize the existing historical data to guide the software development of existing software system development, this paper based on an improved classification and regression tree (Classification and Regression, CART) algorithm software defect prediction models. The paper first principal component analysis of the data predicted correlation dimension (Principle Component Analysis, PCA) between data and reduce the data, and configured according to the theory and optimized CART decision tree algorithm, existing software defect prediction system, and with traditional defect prediction method, the experimental results show that the proposed prediction model has higher prediction accuracy and stability.

This paper proposes a carbon price forecasting system for the participants to quickly and accurately predict the carbon price. The data including the carbon trading price, oil price, coal price and gas price, are first calculated and the data clusters are embedded in the Excel Database by year and season. The Radial Basis Function Neural Network (RBFNN) is constructed in the searching process. The optimal parameters obtained from the RBFNN enable the learning rate parameters to regulate and improve the predicting errors during the training process, enhancing the accuracy and reliability of predictions. By linking the RBFNN and Excel Database, the training stages of the RBFNN retrieve the input data from the Excel Database so that the efficiency and accuracy of the predicting system can be analyzed. Simulation results in this paper will provide an accurate and real-time method for participants to forecast carbon price and raise the market competition in a carbon trading market.

Support Vector Machine (SVM) implicitly maps samples from the lower-dimensional feature space to a higher-dimensional space, and designs a non-linear classifier via optimize the linear classifier in the higher-dimensional space. This paper proposed an improved SVM method based on feature extension and feature selection. The method explicitly maps the samples to a higher-dimensional feature space, perform the feature selection in the space, and finally design a linear classifier with a selected feature set. This paper illustrated the reason why the generalization ability is improved by this technique. The experiment results on benchmark datasets show that the improved SVM greatly decreases the error rate compared with other classifiers, which proves the feasibility of the proposed SVM.

The usual method of outlier analysis is mainly analyzing the outliers according to the Anomaly Index and Variable Contribution Measurement. But in the analysis of large samples of high-dimensional data, this method is difficult. Owing to this, this paper presents a method that weight value for outliers is introduced. The features of outliers are weighted by Analytic Hierarchy Process method. Through this method, the importance of each property of outlier for data mining’s target is rationed, namely the weight number of each property is calculated. And then the correlation values, which represent the degree of relevance between outliers and data mining target, are calculated by using the weight number multiplying by the property value. After correlation values computed, we array the correlation values of outlier from high to low then outlier analysis can become more efficient. At the end of this paper, an instance is presented to demonstrate the maneuverability and feasibility of the method.

In multimodal classification, we look for a set of strategies for mining and exploiting the most informative modalities for a given situation. These strategies are computations performed by the algorithms. In this paper, we propose to consider strategies as advice given to an algorithm by “expert.” There can be several classification strategies. Each strategy makes different assumptions regarding the fidelity of a sensor modality and uses different data to arrive at its estimates. Each strategy may place different trust in a sensor at different times, and each may be better in different situations. In this paper, we introduce a novel algorithm for combining expert strategies to achieve robust classification performance in a multimodal setting. We provide experimental results using real world examples to demonstrate the efficacy of the proposed algorithm.