Groundwater level prediction of landslide based on classification and regression tree

Customer churn prediction is critical for telecommunication companies to retain users and provide customized services. In this paper, a particle classification optimization-based BP network for telecommunication customer churn prediction (PBCCP) algorithm is proposed, which iteratively executes the particle classification optimization (PCO) and the particle fitness calculation (PFC). PCO classifies the particles into three categories according to their fitness values, and updates the velocity of different category particles using distinct equations. PFC calculates the fitness value of a particle in each forward training process of a BP neural network. PBCCP optimizes the initial weights and thresholds of the BP neural network, and brings remarkable improvement on customer churn prediction accuracy.

Particle classification optimization-based BP network for telecommunication customer churn prediction

To ensure the safety of structures, high-speed railway and highway founded on liquefiable soils, liquefaction evaluation is a primary task. Herein, a dynamic set pair analysis method based on variable weights is introduced to assess liquefaction. A concept of connection degree of set pair consisting of the evaluation sample and the classification standard was described to interpret the uncertainty of liquefaction and relationship. Moreover, based on reward and penalty mechanism, state variable vectors were presented to assign composite weights of evaluation indicators in order to take into account impacts of indicator variation. And the integrated connection degree to each grade was calculated to evaluate liquefaction of the evaluated cases. Results from practical example and comparison with field investigation and extension method indicated that this proposed model used to evaluate liquefaction is feasible and effective. In addition, it can quantitatively describe liquefaction hierarchy of evaluated samples.

A novel set pair analysis method based on variable weights for liquefaction evaluation

Soil liquefaction causes extensive damages to buildings and infrastructures during earthquakes. It is very important to identify the significant parameters controlling seismic liquefaction which will enhance the predictability of evaluation models. In order to identify the significant factors, 203 literatures about soil liquefaction are analyzed by bibliometrics, 12 significant factors, such as: magnitude, epicentral distance, duration of earthquake, fines content, particle size, grain composition, relative density, drainage condition, degree of consolidation, thickness of sand layer, depth of sand layer, groundwater table, are identified among 22 influence factors. And then the interrelationship among the 12 main factors was analyzed, and a multilevel structure hierarchy is developed by interpretive structural modeling, which would help understand more clearly about the relationships among these factors. The results show that: magnitude, epicentral distance, duration of earthquake and drainage condition have a direct impact on soil liquefaction, and fines content plays a key role in all factors. The results can provide guidelines for election of factors and the establishment of model in the risk evaluation of soil liquefaction.

Identifying significant influence factors of seismic soil liquefaction and analyzing their structural relationship

Groundwater table often shows complex nonlinear characteristic. Radial basis function (RBF) neural network is increasingly used to predict groundwater table. The traditional RBF training algorithm based on gradient descent optimization method can only obtain the partial/local optimums solution sometimes. Furthermore, man-made selecting the structure of RBF neural network has blindness and expends much time. In training RBF neural network, particle swarm optimization (PSO) algorithm is presented to optimize and automatically determine the parameters and structure of RBF neural network. In order to improve traditional PSO algorithm searching capacity, linear inertia weight and chaos variation operator are presented. Study case shows that, compared with back propagation (BP) or RBF neural network, the new prediction model based on PSO and RBF neural network can greatly increase the convergence speed and precision.

Groundwater Table Prediction Based on Improved PSO Algorithm and RBF Neural Network

Underground water system is a highly complex one. There is a nonlinear relationship between underground water level and its major influential factors. This paper analyzes the shortcoming of BP neural networks and the characteristic of genetic algorithm. On the basis of the hydrologic characteristic, the major factors concerning underground water level are derived, and a dynamic prediction model of underground water level is built, which is used in a specific water resource spot. The case study shows that the model has quick convergence speed and high prediction precision as well as wide practicability in the engineering.

The establishment and application of dynamic prediction model of groundwater level based on intelligent algorithm

Based on matter-element model and the classified analysis theory, an extension clustering prediction method for assessment of seismic liquefaction of saturated sandy soil is proposed. In the method, Eight influence factors on seismic liquefaction, including seismic intensity, epicenter distance, the depth of sand, groundwater level, the blow counts of standard penetration test, mean diameter, coefficient of uniformity, cyclic shear stress ratio, are adopted to establish the classical and limited matter elements, and the dependent function of matter-element and extension set is applied to establish prediction model for seismic liquefaction of saturated sandy soil. The prediction results of the grades of seismic liquefaction of saturated sandy soil can be obtained by means of clustering analysis. The results show that this model can reflect the liquefaction potential of saturated sand objectively and truly; therefore, the extension clustering prediction method is effective and feasible in evaluation of seismic liquefaction of saturated sandy soil.

Extension clustering prediction method for evaluation of seismic liquefaction of saturated sandy soil

The shortcomings of error back-propagation (BP) neural network and the characteristics of genetic algorithm (GA) are analyzed. Based on GA and BP algorithm, the prediction model of vertical ultimate bearing capacity of single pile is built. Real example research shows that the prediction model which based on GA and BP algorithm owns good quality, high precision and is easy to deal with. The method is feasible, effective and has wide applied prospects as well.

Study of prediction method of vertical ultimate bearing capacity of single pile based on genetic algorithm and neural network

Support vector machine(SVM) is a novel and powerful learning method which is derived based on statistical learning theory(SLT) and the structural risk minimization principle.Traditional SVM only deals with the binary classification problems,however,there are large numbers of multi-classification problems in practical engineering.With a view to the main factors with important influence on sand liquefaction,multi-class support vector machines model based on clustering-binary tree is established.The nonlinear relationship between sand liquefaction and influencing factors is obtained from the finite empirical data by SVM.The results indicate that the structure of model base on SVM is reasonable;this algorithm is feasible;and it can predict the sand liquefaction more accurately.

Liu Yong-jian

Papers

The establishment and application of dynamic prediction model of groundwater level based on intelligent algorithm

Extension clustering prediction method for evaluation of seismic liquefaction of saturated sandy soil

Study of prediction method of vertical ultimate bearing capacity of single pile based on genetic algorithm and neural network

Support vector machine model for predicting sand liquefaction based on clustering binary tree algorithm