Research on E-commerce Big Data Classification and Mining Algorithm Based on BP Neural Network Technology

At present, the research on BP neural network has achieved good results in many industries and fields, but there are few projects in the application research of mineral resources mining. Under the social background of the rapid development of electronic information technology, BP neural network and GIS technology are combined to carry out research and application, which will provide a new research path for slope deformation monitoring and disaster prevention in mining area. Therefore, in the paper, the key technology of open-pit mine slope deformation automatic monitoring based on BP neural network and GIS technology was put forward. Firstly, the advantages of BP neural network were analyzed and BP neural network was selected as the prediction model of slope deformation. The artificial fish swarm algorithm was used to improve the BP neural network to improve the performance of the model. Based on the analysis and construction of GIS technology, the combination application of BP neural network and GIS technology was discussed. Through practice, the application effect of the technology was verified, and it has good theoretical and practical value

This paper presents an in-depth study and analysis of the prediction model of force resource recruitment demand using a convolutional neural network combined with a BP neural network algorithm. BP neural network technology is introduced to be applied to enterprise management talent assessment activities. Using BP neural network has strong parallel processing characteristics, as well as unique adaptive learning and feedback adjustment capabilities while combining the traditional enterprise talent assessment system, to build a business management talent assessment model based on BP neural network technology, to circumvent the possible influence of subjective factors in talent assessment, reduce assessment errors, and improve the accuracy and validity of the assessment. The first layer of convolutional layers may only extract some low-level features such as edges, lines, and corners, and more layers of the network can iteratively extract more complex features from low-level features. The constructed applicant reputation evaluation model based on multiplicative long- and short-term recurrent neural network and the hybrid project recommendation model based on conditional variational self-encoder were experimented on Freelancer's dataset for effectiveness, respectively, and the experimental results showed that the proposed employer hiring decision model, reputation analysis model, and applicant project recommendation model have more reliable performance compared with the existing models. The research results achieve more efficient matching of labor supply and demand in the online labor market and provide technical support for the online labor market platform to realize personalized, intelligent, and accurate services for both employers and applicants.

It is difficult to calculate with finite element method or actually measure the temperature distribution of the ceramic sintering at irregular sintering curve. Trained by the temperature distribution data of ceramic sintering analyzed with ANSYS under linear sintering curves including different slopes, the neural network can be used to simulate that under irregular sintering curve at certain precision, so the temperature evolution of the ceramic hot geometry centroidal point (HGCP) can be fast obtained by the result simulated with the trained neural network. This research sets up series-parallel BP neural network model with MATLAB. The ceramic sintering date analyzed with ANSYS under linear sintering curves with ten different slopes is used as training sample of the neural network which is tested by the sample under non-linear sintering curve. The result indicates that the BP neural network is feasible for simulating temperature distribution of the ceramic sintering.

To further study the reliability of human factors in dam break accidents, by establishing an index system of human factors influencing factors, human errors are divided into four categories and twelve sub-categories of "people, technology, organization, and environment". Combined with BP Neural network technology, we input more than 60 historical dam failure data to get the weight of human error under different dam failure reasons.

In view of the deficiency of current gas monitoring systems in coal mine roadway excavation, a two-level information fusion technology, which adopted the adaptive weighted algorithm and the BP neural network technology, was applied to gas monitoring. The results show that the adaptive weighted algorithm can realize self-regulation by decreasing the weight value of the failed sensor automatically, so as to eliminate the effect of the failed sensor and ensure the effectiveness and accuracy of the gas monitoring system. The BP neural network can not only effectively predict the gas gush quantity of the excavation roadway, but also accurately calculate the gas concentration in the region where one or more sensors have failed, so as to provide the basis for judging the safety status of the roadway excavation. The experiments prove the superiority and feasibility of the application of information fusion in gas monitoring.

The research focuses on using BP Neural Network Technology and Principal Component Analysis to prevent debt crisis. Using AI (artificial intelligence technology) to establish an implicit debt prediction model is a new topic. This paper establishes an effective indicator system, uses principal component analysis (PCA) to extract comprehensive indicators, then uses the BP neural network model to derive the predicted value of the local government implicit debt in the past five years, and finally compares the predicted value with the estimated value. The results show that in 2017, 2019, and 2020, the estimated implicit debts of local governments did not reach the warning level, and the risks were controllable. In 2016 and 2018, it exceeded the warning forecast value, which is worthy of attention. This paper serves as an example to be extended and applied to the construction of early warning system in various places.

Abrasive water jet can be applied to perforate the oil formation rock. The perforation depth generated by abrasive water jet is nonlinearly influenced by so many factors that it is difficult to mathematically correlate the perforation depth with influencing factors. So the back propagation (BP) neural network is introduced to establish the model for predicting perforation depth generated by abrasive water jet. Firstly the fundamentals of BP algorithm are briefly reviewed in this paper. Then regarding the special application of BP network in this research, the methodology of how to select sample set, how to optimize the BP hierarchy and the number of nodes in hidden layer is given in detail. The established BP model is trained and tested by an experimental data set. The test results show that the prediction precision of the BP model can completely meet engineering requirements with the average relative prediction error of only 3.54%.

Sports Event Data Acquisition Method Based on BP Neural Network and Wireless Sensor Technology

This paper puts forward a kind of evolutionary algorithm and the neural network combining with the new method of optimization of hidden layer nodes number of particle swarm algorithm of neural network. The BP neural network technology is a kind of more mature neural network method, but there are easy to fall into local minimum value, unable to accurately determine the number of hidden layer nodes of the network, the disadvantages such as slow convergence speed. This paper puts forward the optimization with hidden node number of particle swarm neural network (HPSO neural network) is the hidden layer of BP network node number as a particle swarm optimization (PSO) algorithm is an important optimization goal, network of hidden layer nodes and the number of each BP network weights and closed value together, common as particle swarm algorithm optimization goal.

This paper describes an experiment on Parkinsons disease classification using multiple classification algorithms for comparison. Parkinsons disease is a common neurological disorder, and early diagnosis and classification are important for the assessment of treatment and prognosis. Therefore, the research implications of this paper are clear. The classification algorithms used in the experiment include adaboost classification model, XGBoost classification model, logistic regression regression model, random forest plain Bayesian classification model, bp neural network and support vector machine. The experimental results show that adaboost classification model performs well when dealing with small sample data, XGBoost classification model performs well when dealing with large-scale datasets, and logistic regression regression model and random forest plain Bayesian classification model also have good performance. The bp neural network and support vector machine, on the other hand, perform poorly in terms of classification results and require a much larger dataset for support. These experimental results have important reference value for the classification and diagnosis of Parkinsons disease. Different classification algorithms are suitable for different dataset sizes and characteristics, so in practical applications, we can choose different classification algorithms according to the size and characteristics of the dataset to achieve the optimal classification effect. In conclusion, the results of this paper provide a reference for the classification and diagnosis of Parkinsons disease, as well as a guide for choosing appropriate classification algorithms. In the future, we can further expand the dataset size and use more classification algorithms for comparison to improve the accuracy and robustness of Parkinsons disease classification.

With increased internet usage, online transactions have been on the rise. One of the most prevalent problems faced is credit cards frauds. While web applications and mailing services are heavily spammed, the upsurge of handheld mobile devices has led to an outburst of heavy mobile credit card spamming. The matter is more severe in mobile devices due to lesser sophisticated filtering mechanisms in built in mobile operating systems. Recent advancements in electronic commerce and communication systems have significantly increased the use of credit cards for both online and regular transactions. However, there has been a steady rise in fraudulent credit card transactions, costing financial companies huge losses every year. The development of effective fraud detection algorithms is vital in minimizing these losses, but it is challenging because most credit card datasets are highly imbalanced. This work proposes a supervised machine learning algorithm to be trained to detect credit card frauds based on the Bayes Net with penalty based regularization. It is shown that the proposed approach attains higher classification accuracy compared to existing work. Keywords— Credit Card Fraud Detection, Machine Learning, Feature Selection, Imbalanced Datasets, Probabilistic Classifier, Classification Accuracy

We take a well log in petroleum exploration and development as an example of the big data mining, and adopt three regression and two classification algorithms: the multiple regression analysis (MRA), the error back-propagation neural network (BPNN), the regression of support vector machine ( R -SVM), the classification of support vector machine ( C -SVM), and the Bayesian successive discrimination (BAYSD). It is well known that MRA, BPNN and R -SVM are regression algorithms while C -SVM and BAYSD are classification algorithms, and only MRA is linear algorithm whereas the other four algorithms are nonlinear algorithms. From this case study, we can draw the following five major conclusions: a) Since C -SVM is the best classifier, it is employed as a data cleaning tool. b) Since MRA is a linear algorithm, its total mean absolute relative residual R * (%) can express the nonlinearity of studied problem. For this case study, R * (%)=52.14 showing the nonlinearity of the studied problem is strong. c) Since both MRA and BAYD can establish the order of dependence between a dependent variable and independent variables, each of MRA and BAYD could serve as a pioneering dimension-reduction tool in data mining. In the case study, since the nonlinearity of the studied problem is strong, the nonlinear algorithm BAYSD can serve as a pioneering dimension-reduction tool, but the linear algorithm MRA cannot. d) Since the nonlinearity of the case study is strong, BPNN and R -SVM are not applicable though they are nonlinear algorithms, whereas other two nonlinear algorithms C -SVM and BAYSD are applicable, indicating the nonlinear ability of C -SVM and BAYSD is higher than that of BPNN and R -SVM. e) Comparing the two applicable algorithms C -SVM and BAYSD for this case study, it is seen that R * (%) of C -SVM is less than that of BAYSD; BAYSD can serve as a pioneering dimension-reduction tool, but C -SVM cannot; it is easy to code the BAYSD program whereas it is very complicated to code the C -SVM program, so BAYSD is a good software for this case study when C -SVM is not available. Key words : Big data mining; Well log; Data cleaning; Dimension-reduction; Regression; Classification

To alleviate the pressure of data size, data transmission and data processing in the huge data dimension of the Internet of things., data classification is realized based on back propagation (BP) neural network algorithm. The working principle is deduced in detail. For the shortcomings of slow convergence and easy to fall into the local minimum, the combination of variable learning and momentum factors is used to improve the traditional back propagation algorithm. The results show that the optimized algorithm improves the convergence speed of the network to a certain extent. Therefore, it is concluded that the back propagation neural network has higher classification success rate when classifying multidimensional data in Internet of things.

BBO-BPNN and AMPSO-BPNN for multiple-criteria inventory classification

In order to solve the problem of identifying ship trajectories with different motion patterns in waterways, a ship trajectory classification algorithm based on machine learning techniques is proposed. First, the area of waterways is divided and the trajectory selection rules are formulated to construct the label matrix. Then, the piecewise cubic Hermite interpolation algorithm is used to extract the feature points of the original trajectory data from the perspectives of the equal time interval in recording time and the equal space interval in spatial distribution to construct the trajectory feature matrices. Finally, in order to train and test the classification model, the label matrix and trajectory feature matrices are put into the model and the parameters are optimized. The AIS trajectory data in Huaian section of the Beijing-Hangzhou Canal are selected for the trajectory classification experiment. In this research, the broad learning system (BLS), the back propagation neural network (BPNN) and the support vector machine (SVM) are chosen as the machine learning classification methods. The results show that the trajectory classification model based on BLS is superior to those based on BPNN and SVM in classification accuracy and training time.