The Discussion Of Several Kinds Of Approximation Algorithms And The Prediction Of Ceramic Art Price

In recent years,soaring ceramic art prices has become conditional factors of art ceramics market development in our country,and the scholars have not been deeply through the quantitative research on the affect ceramic price in today’s society.This article uses three different approximation algorithms to forecast the ceramics price respectively deeply.The first algorithm is based on transform weight shared value spline interpolation prediction algorithm for study the ceramics price projections,it’s mainly from the 10 factors of affecting the price of ceramic art to analysis,we can research from the 10 affect factors that brought about the high-dimensional data,in order to overcome the weight of the whole processing algorthm’s inadaptability,so we put forward the based on transform’s weight shared value spline interpolation prediction algorithm.This algorithm can according to the different correlation coefficient between the influence factors and weights,then through the transformation,normalized and the weight of shared value,and then according to spline interpolation to structure the weight function to predict all of the new sample’s weight value.The example shows that the algorithm is better than other algorithms in precision,and it has good stability.The second algorithm uses T-S fuzzy neural network algorithm,the algorithm is based on the affection of ceramic art price’s statistical average method and the transformation,then get the value of each factor in each data and use quantitative to measure the qualitative data.Through this transformation,we use the T-S fuzzy neural network to train the data,so we can get the network parameter,and then through the simulation to predict other testing data prices.The example shows that this algorithm has better generalization ability,and has better prediction precision,the error is less than 9%.The third algorithm uses the associated with variable weight’s wavelet neural network algorithm to research the forecast of ceramic price,this algorithm is based on the compactly supported of wavelet function and the analysis of smoothness,then combined with the neural network algorithm,at last we put forward a kind of based on the associated with variable weight’s wavelet neural network algorithm.This kind of algorithm is more applicable for weights of smaller,the prediction error is almost 0 everywhere,and has good generalization ability.Through these three algorithms,we are mainly from the quantitative of factors that influence the price of ceramic to study the forecast of ceramic art price deeply,the study found that these three experimental results of the algorithm’s error is very small,and the precision is more higher,too.In the future research,we can improve these three kinds of algorithms so that wecan make the experimental error minimum,and the accuracy is the most highest.