Search

Knowledge community detection algorithm is a main technical means to knowledge discovery in complex network. It is used to discover the potential community structures to provide a basis for web services, for example, personalized recommendations. In this paper, a parallel algorithm for knowledge community detection based on MapReduce is proposed. This paper, through various experiments, proves that the efficiency of the parallel algorithm is increased significantly.

This paper focuses mainly on a clustering method for pruning Fully Connected Backpropagation Neural Network (FCBP). The initial neural network is fully connected, after training with sample data, a clustering method is employed to cluster weights between input to hidden layer and from hidden to output layer, and connections that are relatively unnecessary are deleted, thus the initial network becomes a PCBP (Partially Connected Backpropagation) Neural Network. PCBP can be used in prediction or data mining more efficiently than FCBP. At the end of this paper, An experiment is conducted to illustrate the effects of PCBP using the submersible pump repair data set.

The main objective of this investigation is to explore new similarity algorithms of staff similarity in technology innovation team. First, this paper proposes the knowledge representation model of technology staff based on network, and the cliques after clustering according to network feature expresses the sub-fields. Second, from the view of knowledge contained in technology staff, this paper proposes the similarity algorithm based on VSM and the similarity algorithm based on sub-field. Finally, we use the staff classification of one technology innovation team as case study. The experiment results reveal that the similarity of the new methods is accurate than that of the old method, and the information obtained by the new methods is more than that obtained by the old method.

This paper discusses a new kind of mode in database design——mode of “column changes line”. This paper discusses the definition of the mode, and in detail analyses the mode through specific cases, and summaries its specific advantage and applied scenes. In the end of the paper, it discusses the relationship between the mode and four-atomization theories in database design.

Datamining used by an enterprise is an powerful tool to provide quickly our enterprise useful information and improve well our decision making . Based on clementine, the paper has given a solution of datamining, has certain innovation, higher theoretical significance and practical value. This paper mainly researches how to build a datamining application based on clementine.

The knowledge database of ready-to-wear included the clothing knowledge required in the process from design to manufacture. It is to carry on management and storage of the original trivial clothing knowledge, achieve rapid information access, delivery and response. People hope garment fit their body better, but now garment is produced in bulk. The challenges the company faces is to achieve intelligent pattern making and pattern individuation based on the knowledge database of ready-to-wear.to improve the production efficiency. In this paper, the relation among the different data in the knowledge database of ready-to-wear are studied. The key is how to achieve transforming the human body data, style and craft data to the pattern plate data, and ascertain the constraint relation between body measurement data and pattern plate data. Man’s shirt is taken as an example in this paper. The conversion function and the fitting curve of the back armhole are obtained through analysis of the pattern plate data in order to provide the foundation for achieving the transformation from data information to pattern sketch.

To discover requested web services in Internet, this paper proposes a new web service discovery algorithm based on ontology. It uses tree-form data structure to describe the web services and give all the nodes a weight value by certain strategy, then compute the semantic similarity between the web services requested and the services registered. To validate the feasibility and effectiveness of the algorithm, the paper constructs a self-developed prototype system to show how well it works. The experiments prove that this algorithm has high recall rate and precision than other methods.

This paper presents a power load prediction mathematic model based on wavelet neural network theory which is a new neural network model thrives from wavelet analysis theory. By applying wavelet analysis, this paper can analyze power load in any dimension and collect it effectively. At the end of this paper, the authors analyze the model parameters and discuss various factors deeply.

In text association classification research, feature distribution of the training sample collection impacts greatly on the classification results, even with a same classification algorithm classification results will have obvious differences using different sample collections. In order to solve the problem, the stability of association classification is improved by the weighing method in the paper, the design realizes the association classification algorithms (WARC) based on rule weight. In the WARC algorithm, this paper proposes the concept of classification rule intensity and gives the concrete formula. Using rule intensity defines the rule adjustment factors that adjust uneven classification rules. Experimental results show the accuracy of text classification can be improved obviously by self-adaptive weighting.

This paper mainly introduced the basic theory of Hidden Markov Model (HMM) and Support Vector Machines (SVM). HMM has strong capability of handling dynamic process of time series and the timing pattern classification, particularly for the analysis of non-stationary, poor reproducibility signals. It has good ability to learn and re-learn and high adaptability. SVM has strong generalization ability of small samples, which is suitable for handling classification problems, to a greater extent, reflecting the differences between categories. Based on the advantages and disadvantages between the two models, this paper presented a hybrid model of HMM-SVM. Experiments showed that the HMM-SVM model was more effective and more accurate than the two single separate models. The paper also explored the application of its database system development, which could help the managers to get and handle the data quickly and effectively.

According to the companies' purified water testing data, this paper established a product Boolean test model and used principal component analysis for analyzing security risk of purified water, and then based on probability theory, this paper set the distribution of testing batches to help the inspection department attaining the optimal sampling plan in the limited funding inspection.

Ontology mapping is a crucial task to enable interoperation and interoperability between related ontologies. Combining multiple mapping strategies can improve the mapping accuracy in ontology mapping. However, how to give weight values in multiple strategy ontology mapping is very difficult. Aiming at the problem, characteristics based adaptive weighting method is proposed in the paper. Six similarities is selected to reflect different kinds of ontological information. Ontologies characteristics related to the six similarities, which include self-characteristics and mutual-characteristics of both ontologies, are used to describe both ontologies to be mapped. Weight of each similarity is specified according to the characteristics of both ontologies. Experiment results showed that the method in the paper can get a better effect than traditional weights combining methods.

In this paper, we present a solution developed at rough set theory and weight value to make multiattribute decision with incomplete information. The paper defined the concepst of breaking points and recovered the incomplete information system according to the relationship between the condition attributes and decision attributes. And introduced OWGA(ordered weight geometric averaging) operators to calculate the aggregation value of each project. Finally, selected the project with the maximum aggregation value as the best decision making. The illustration and experiments were implemented and the results indicate that the method is effective and efficient.

In order to solve the problem in k-means algorithm that inappropriate selection of initial clustering centers often causes clustering in local optimum and the time complexity is too high when handling large amounts of data, a fusion clustering algorithm based on geometry is proposed in this paper. The result of experiments shows this algorithm is better than the traditional k-means and the k-means++ algorithms, with higher quality and faster speed. And at last in this paper, we apply it in marine engineering.

Decision tree classification is one of the most widely-used methods in data mining which can provide useful decision-making analysis for users. But most of the decision tree methods have some efficiency bottle-necks and can only applied to small-scale datasets. In this paper, we present an new improved synthesized decision tree algorithm named CA which includes three important parts like dimension reduction, pre-clustering and decision tree method, and also give out its formalized specification. Through dimension reduction and synthesized pre-clustering methods, we can optimize the initial dataset and considerably reduce the decision tree’s input computation costs. We also improve the decision tree method by introducing parallel processing concept which can enhance its calculation precision and decision efficiency. This paper applies CA into maize seed breeding and analyzes its efficiency in every part comparing with original methods, and the results shows that CA algorithm is better.

The flood of spam promotes the development of anti-spam technology. In this paper, we bring forward the Bayesian filter technology based on the distributed architecture, which can realize the sharing of the Bayesian learning outcomes among servers within the system, so as to increase the accuracy of spam recognition. We, in the paper, discuss the sharing model of information with spam features under the distributed architecture and the spam identification process; analyze the Bayes algorithm and carry out the relevant improvements; design the Bayes Filter based on distributed architecture on the above basis and verify the effect of the filter by experiments.

Based on the idea of the neural network, intelligent computing methods are used to analyze temporal and spatial data. We present the temporal and spatial autocorrelation moving average (STARMA) model based on the in-depth systematic study on time sequence of hybrid model. Firstly this paper uses radial basis function neural network to extract the temporal and spatial sequence which is non-stationary caused by large-scale non-linear trend, secondly this paper presents STARMA modeling of small-scale random spatial and temporal variation. Comparative analysis between the original data and the forecasting data shows that proposed hybrid model has better performance of fitting and generalization.

This paper presents a comparative study of clustering using Artificial Intelligence (AI) techniques. There are 3 methods to be compared, two methods are pure method, called Self Organising Map (SOM) which is branch of Artificial Neural Network (ANN) and Genetic Algorithm (GA), while one method is hybrid between GA and SOM, called GA-based SOM. SOM is one of the most popular method for cluster analysis. SOM will group objects based on the nearest distance between object and updateable cluster centres. However, there are disadvantages of SOM. Solution quality is depend on initial cluster centres that are generated randomly and cluster centres update algorithm is just based on a delta value without considering the searching direction. Basically, clustering case could be modelled as optimisation case. The objective function is to minimise total distance of all data to their cluster centre. Hence, GA has potentiality to be applied for clustering. Advantage of GA is it has multi searching points in finding the solution and stochastic movement from a phase to the next phase. Therefore, possibility of GA to find global optimum solution will be higher. However, there is still some possibility of GA just find near-optimum solution. The advantage of SOM is the smooth iterative procedure to improve existing cluster centres. Hybridisation of GA and SOM believed could provide better solution. In this study, there are 2 data sets used to test the performance of the three techniques. The study shows that when the solution domain is very wide then SOM and GA-based SOM perform better compared to GA while when the solution domain is not very wide then GA performs better.

In this paper, the usage possibilities of personal statistics are introduced, which can be applied to improve the patient-specific evaluation in health monitoring systems. The aim of these techniques is to obtain reliable results based on previous measurements. This goal can be achieved by membership function tuning or modification, as well as by a pre-processing method, which is used to judge whether a situation is normal or not. In the latter case, a further requirement, that the appropriate result should be available in time, can also be fulfilled. If the situation is judged to be critical then a reduced model is evaluated instead of the full one.

In this paper, a new filter network is presented that is based on Radial Base Function Networks (RBFNs). The output layer of the network is modified, in order to make it more effective in certain fuzzy control systems. The training of the network is solved by a clustering step, for which two different clustering methods are proposed. The suggested structure can efficiently be used for data classification.

Scientific and normative evaluation index system of smart distribution grid can provide effective guidance and advice for planning and construction of smart distribution grid. Based on the level and target of smart distribution grid construction, combining with the current situation and the development of smart distribution grid in China, this paper built a preliminary framework of evaluation index system of distribution grid intelligent level by hierarchical method, the system contains base layer, measurement and control network layer, management layer and effect layer. The key and new indicators are defined for each layer. We can estimate the distribution grid intelligent level objectively and comprehensively, and a reference for improvement and development of smart distribution grid will be provided with the help of this evaluation index system.

In order to evaluate rational buoys layout of the route quantitatively, the method of fuzzy comprehensive evaluation is applied to establish a rational buoys layout evaluation model. According to the 9 main factors that affect rational buoys layout, a quantitative assessment model is established. The AHP is used to determine weighted coefficients of assessment index. In the paper, the assessment for the buoys layout of three routes is carried out.

Information extraction of object oriented method was applied to study the land use change, this paper puts forward a kind of application of object oriented remote sensing image classification to extract land change information extraction methods to change the land use, introduces the calculation of object oriented classification, classification of ideas and technology process, automatic extraction of water information in the process of classification to discuss the specific computer.

3D quantitative prediction can be summarized as finding the combination parts of favorable metallogenic information based on the 3D geological models and cubic block models. Based on metallogenic prediction theory, relying on 3D visualization technology, 3D database technology and statistical calculations, this paper established the technical processes of 3D quantitative prediction and evaluation of deep mineral resources which including 3D geological modeling, prospecting model establishing, mineralization favorable information analysis and 3D quantitative prediction and evaluation.The favorable metallogenic information analysis and extraction which implemented based on 3D cubic block models extended the prospecting method from 2D to 3D space, and realized the visualization of deep quantitative geological information from the 3D point of view. The method of using 3D spatial exploration flag variable to realize 3D prediction of deep concealed ore provides a new way of prospecting prediction study of deep mineral resources.

With the development of global economy, logistics industry is gradually developing. Along with the rapid development of e-commerce, Third Party logistics suddenly emerges, which has promoted the development of logistics industry. With the development of logistics industry and information technology, traditional logistics management pattern has already not adapted to the need of modern logistics management. Faced with a lot of product information and customer relationships, new modes of information management must be established to improve the efficiency of management. This paper mainly discusses the application of basic data mining technology in the logistics enterprise management.

This paper presented a new method to evaluate Remote Sensing image quality, by comparing ZY1-02C, ZY3, and SPOT5 images on the engineering quality and spectral quality. It is important to explore new options to evaluate different Remote Sensing image sources quality, in order to ensure the users could apply a best fit data source to environmental monitoring, ecological monitoring and so on. In this article, there were three aspects in the engineering quality assessment part, including the statistical character, the texture and the energy. And in the spectral quality assessment part, the imaging space, the curve space and the characteristic space were built to compare and measure different spectral ability of extracting ground objects among ZY1-02C, ZY3 and SPOT5 images. The result shows such a Remote Sensing image quality assessment can be generalized to choose suitable data source for some specific field.

This paper introduces the method of Projection Pursuit the idea, principle, modeling steps and application of the model in water quality evaluation of Dehuixin river. Use the Real-code Accelerating Genetic Algorithm to optimize the problem and then comparing the result with other results of Pollution Index Method and Fuzzy Evaluation Method. It shows that the method of Projection Pursuit can better evaluate the water quality, its result has higher objectivity and authenticity through the calculation, analysis and comparation.

The existing sample selection methods are inadequate in the determination of defect sample set, and lacking necessary metrics or evaluation methods of the defect sample set. This paper determines the typical defect sample set according to the selecting principles and proposes a metrics method of defect sample set to measure whether the defect sample set is representative of the defect library. At present, defect sample selection mainly depends on brainstorming and expert experiences in the electric system. However, for those organizations which have no experiences orfor large databases, there are significant difficulties on sample selection. The proposed defect sample selection framework and measurement criteria can effectively improve the situation and ensure that the selected sample set can well represent the defect library.

On the basis of six kinds of air pollutant data provided by the Xi’an Environment Protection Bureau, Two kinds of air assessment model were presented in this paper. Firstly, air quality index (AQI), which has been adopted as a part of national standard in China, was used to assess the air quality of Xi’an in 2013 winter. We also introduced a fuzzy self-organizing feature map (SOFM) model to classify air quality in an unsupervised and comprehensive way. As a further research, a hybrid model was put forward based on Evolutionary Strategy (ES) and SOFM. With SOFM neural networks embedded into ES, the sensitivity of SOFM neural networks to the initial weight matrix and sequence of exemplar input is overcome by the global optimization of ES. The results of our work demonstrate that the ES-SOFM Hybrid Model is quite appropriate techniques for air quality assessment. Unlike AQI method, SOFM’s result is decided by all pollutant instead of only the most serious one. No matter what kind of method, all assessment results show the very serious air pollution in Xi’an. The government and every citizen must take steps at once to prevent air quality from further depravation.

A comprehensive analysis on the statistical data about the energy, environment, population and development of the social economy, is the necessary condition to solve regional environmental and social problems and to realize the sustainable and scientific development. But different statistical unit is the main obstacle of the integrated application of statistical data. Practice show that statistical data spatial gridding is an effective way to solve this problem. This paper presented the factor-based statistic spatial gridding process for the sum variable. This method firstly determined the factors that affecting the spatial distribution of statistical objects. Then it used the AHP method to obtain the weight of each factor, and applied the weights to grid calculation. The experiment proved that this method made the grid result more realistic.

The aim of this paper is to forecast short term variation of the regional wind energy resource based on the data captured from several wind measuring stations. Firstly, the main spatial patterns are extracted by SVD (singular value decomposition) method, and then the time coefficient series corresponding to principal spatial patterns are processed and forecasted by SVM (support vector machine). Furthermore, according to the SVD method, the new forecasted time coefficient series are used to inversely calculate the wind speed in the future. Finally, the validity and performance of this forecasting method is tested in case study.

Inthis paper, according to the more important ten evaluation indicators, the fourgrades ideal evaluation is established corresponding to the level of healthstate of bridges. Combined with associative memory capacity of discreteHopfield neural networks, a new health state evaluation of bridges ispresented. Five bridges is evaluated by the model, the network connectionweights is obtained by iterative learning using the outer product method. Thesimulation results shows that the health evaluation model can evaluate thehealth state of bridges fast, accurately and intuitively.

In wireless sensor networks, traditional link selection algorithm needs lots of data packages as testing samples, but the nodes of WSN are battery-powered, so the energy is extremely limited. To overcome this shortcoming, the aim of this paper is to propose three new link selection algorithms based the concept of Bayesian approach. Simulation results demonstrate that the three algorithms based on Bayesian approach have a higher success rate than empirical-algorithm by about 10 percent in selecting the highest quality link with the case of small samples. Among them, BSLA-EB has a good adaptability and it can get better experimental results.

With the rapid development of computer networktechnology, and also can be used as well as the increasing number of users ofcomputer network, how to effectively guarantee the network information securitybecomes a key technology of computer network. This paper through the literatureinvestigation method to build the information system security risk evaluationindex system, and combining the expert evaluation process could not get all theinformation in the information system security, a method of information systemsecurity risk assessment is put forward. Was carried out on the key technologyof information system based on big data discussion and analysis, and brieflyintroduces its development prospects. Secondly illustrates the necessity ofintrusion detection, and the concept of intrusion detection and the model isgiven. Finally summarizes many kinds of intrusion detection method and thesystem structure, discusses the existing problems of this field and theresearch direction in the future.

The aim of this paper is to propose a new information fusion method for the problem of multi-sensor target recognition. Multi-sensor information fusion problem contains many characteristic indexes, and thus it can be regarded as a multi-attribute decision making problem. The new fusion method is put forward based on preference selection index method. The new information fusion method is not necessary to assign relative importance between attributes, but overall preference value of attributes are calculated using concept of statistics. Thus the new method can overcome the subjective randomness of subjectively weighting method. An applied example proves that the method is both effective and exercisable.

Intelligent logistics is the use of integrated intelligent technology, which makes the logistics system to mimic human intelligence with the thought, perception, learning, inference and solve some problems of logistics in their ability. Association rule mining is usually more applicable and recorded in the index of discrete values. This paper analyzes theory and algorithm research of association rules data mining and presents design and development of intelligent logistics system based on data mining and association rules technology.

Vessel survey is a kind of technical method to ensure the health and safety of the vessel. The main purpose of the paper is to find a way to extract workflow of vessel survey field operation (VSFO) directly from rule-based document (RBD) of the field, introduces the relevant researches on the vessel survey field, and analyzes the novel way to improve the quality of vessel survey via intelligentized aided technologies. The establishment of VSFO workflow is treated as the major concern in the work, and relative studies in other fields are introduced. After the comparison, three characteristics of VSFO workflow are discussed. On the basis of previous studies, a semantic approach of the establishment is proposed, and the whole process consists of several steps. Firstly, the structure of RBD in vessel survey is discussed; meanwhile, vessel survey domain ontology (VSDO) is built with rules in RBD. Secondly, the classification algorithm is designed with the original framework provided by domain experts. Finally, the distance calculation of clustering is divided into three parts: reachable distance, structural complexity and survey content complexity.

Smart sensor has the following three advantages: realize the information acquisition of high precision and low cost, it relates to the micro mechanical and microelectronic, signal processing and computer technology. The purpose of data mining is to discover knowledge. Knowledge of association rules mining aims to find out the related information hidden in the database. The paper presents application analysis of smart sensor node based on data mining association technology. Experimental results show the proposed methodology has advantages in the management of the intelligent node.

Aiming at the problems existing in maintenance of flight simulator, considering characteristics of expert system, the architecture and fault diagnosis methods of flight simulator, the structure and key realization technology for flight simulator fault diagnosis expert system are studied; the flight simulator fault diagnosis expert system design proposal is completed. Research on transforming from fault tree pattern to rule and diagnosis knowledge representing are emphasized, expert system knowledge base is constructed by means of frame plus production knowledge representing method, then certainty and rule based backward chaining inference engine with depth-first and heuristic search mechanism is designed, the system diagnosis flow and inferring flow are also developed in this paper. Finally the expert system is tested by a fault solving demonstration.

In the era of big data, due to the rapid expansion of the data, the existing incremental text clustering algorithm has the drawback that the efficiency of algorithm will sharp decline with the time and data volume increasing. Because of poor timeliness and robustness, the algorithms are hard to be applied in practice. In this paper, we propose a distributed model framework of Single-Pass algorithm based on MapReduce, the experiments result of increment text cluster is accuracy, the algorithm effectively improve the computing efficiency of the algorithm and real-time of result. Algorithm has a great prospect under the background of big data.

Machine aided human translation (MAHT) for the abstract of patent texts is an important step to the deep processing of the patent data, where the terms have significant application value. This paper investigates the automatic term recognition (ATR), and proposes a new hybrid method based on two-phase analysis and statistic to generate English candidate terms. The segments including stop words were not simply discarded; instead, a rewriting method using beginning patterns, ending patterns, and inner patterns on the phase two was employed for the processing of the segments. In the meantime, generalized statistical measures were used for the evaluation of the candidates such as the generalized mutual information (MI), Log-Likelihood Ratio (LLR), and C-value to filter the low score’s candidate terms and to attain the intersection set of them. The experiments on the patent abstract texts extracted randomly show the availability of the method.

With the development of sports, sports e-government platform applications more widely. A number of schemes on how to e-government platform of sports to make decisions, to choose the most appropriate solution is always facing the problem for policy makers. This paper uses linear weighted comprehensive evaluation method for comprehensive evaluation of the program. First, based on the former research results, construct the evaluation index system and weight; Then, follow the steps for solving a mathematical model, including a single index evaluation, the first level indexes and the total target evaluation; Finally, based on a mathematical model three programs to evaluate e-government platform, to choose the optimal scheme. The results show that the research method has clear ideas, the calculation is simple, easy to operate, objective results and other characteristics.

Maritime search and rescue is important for the maritime traffic safety. A maritime search and rescue capability evaluation algorithm based on cloud model is proposed in this paper. Firstly, we established the evaluation system; Secondly, we determined the expert evaluation cloud model according the scoring; Thirdly, we calculated the evaluation result by the comprehensive cloud; Finally, we used the digital characteristics to restore the cloud chart. The experimental result shows that our method is objective and accurate.

The idea of using time series analysis to predict product is not new. At present, there are many food resources prediction and evaluation, and in the agricultural research it is prediction good method of international popular food like method, remote sensing prediction method, multiple regression and other methods to predict yield. In this paper, the main production data is used in Jilin province in recent decades, the application of time series analysis method of comparative study, in order to serve agriculture and choose a more accurate prediction of time series model in recent years of maize yield in Jilin province. The results showed that ARIMA (2,1,1) could correctly simulate and forecast the products of maize, providing an important method for accurately predicting the yield in formation of agricultural products.

This paper introduces the ontology as the basic description of the domain feature model and defines the description methods of static semantics and behavior protocols. On this basis, it proposes the Architectural Pattern based on the feature model and searching and matching methods of the Design Pattern. Variable point information on the domain feature model could guide the specific application-oriented architecture to cutting and customization so as to lay a solid foundation for the establishment of trace relationships and the system evolution management based on feature tracking.

he information technology has greatly driven the development of Smart Grid, however, the risk of system operation has been significantly increased simultaneously with improving the intelligent operation level of Grid system, therefore, how to construct an efficiency evaluation model of electric power information system effectively has been becoming an important subject to be explored for many experts. And based on that, this paper presents an efficiency evaluation model of electric power information system in view of the adjacency matrix. Firstly, the attack process is divided into two stages from the perspective of an attacker, namely, selecting to attack the access points and attacking other access points; secondly, it describes the intelligent selection capability of the attacker by using the transfer probability index between the nodes; finally, it conducts the computational analysis by introducing an example, to prove the practical applications of the proposed methods.

This paper describes two methods for the industrial quality inspection: Supervised classification algorithm Chi-Square Automatic Interaction Detector (CHAID) and unsupervised clustering algorithm Self-Organizing Map (SOM). The classification and clustering are modelled in IBM software SPSS. Models’ functioning is illustrated on a wheel assembly geometric features inspection. The classifying accuracies are compared for the two methods. CHAID has shown better classifying ability than SOM, while SOM can be used to improve quality of predictor values, and therefore classifiers accuracy.

In order to make up the real-time performance of tracking and control information database, this paper design a kind of two-layer’s real-time data storage model based on memory database and relational database. In this article, the two-layer’s real-time data storage mechanism and life cycle are expounded in detail, analyzing and inducing the real-time data characteristic and storage strategy, putting forward the memory database’s self-adaptive index algorithm of T-tree index and hash index, and introducing the database synchronization mechanism between the memory database and relational database and so on. In this way, so as to improve and optimize the real-time, reliability and security of database, provides a reliable data guarantee for future expansion of the real-time application.

Lead, Mercury and Cadmium etc as the main evaluation index of heavy metal pollution established relational data model. The rough set theory is introduced, use the existing algorithm (Combinatorial Completer) to fill the missing value. After data preprocessing, use the DBMAS algorithm proposed in the paper to calculate the important degree of heavy metal pollution factors, in order to provides a more objective evaluation index weight for evaluation of heavy metal pollution.

The paper establishes the objective functional model of sheet metal forming in drawing process with ANN, a mapping between sheet forming parameters and performance evaluation indexes was built, which provides important preferences for researching and optimizing these parameters. It obtains neural network model of high precision through the training of cross experiment method. At last a model was built. According to the test results, the error of the network were less than 5%.That means the network is available, and also it establishes foundation of the process parameters optimization.