A Study Of Using Statistical Simulation Technology To Improve The Goodness Of Fit Of The Model

As early as the end of the 17th century,people already know the frequency of event occurrence to determine the probability of events.Development of the computer makes possible to achieve large-scale random sample tests.Computer simulation of random phenomena,technology,applied to the statistical study,analyze and use the results to practical problems,there is a computer simulation.Statistical simulation can be constantly repeated,changed the structure and parameters of the system are relatively easy,both relatively realistic simulation of reality,experiments and their relatively low cost,as well as shorter working cycles,improve work efficiency.This article is,in this context,economic forecasting using statistical simulation technology in the study of the typical problems of the exploratory study.Based on the production cost forecast model,for example,first use a descriptive analysis and contingency table analysis method of geological factors on the production cost of special effects,when in a certain range of geological factors change does not affect the oil production costs change significantly only when it exceeds a certain critical value,will have a significant impact on production costs.According to these factors the data features three simulation experiments were conducted using statistical simulation technology,and comparative analysis of the results of simulation,simulation test conclusion is given.Finally,the use of traditional models and improved models of the 78 samples of oil field data modeling and parameter estimation,and McKinnon J tests two models,comparative advantages and disadvantages of the two models.Empirical results show that:showing interval changes in data,index ordering can improve the fit of the regression model index,and prediction accuracy can be significantly improved.In a certain range,data distribution more scattered,greater fluctuations,the model goodness of fit index ordering improvements more obvious.But not interval changes as the main characteristics of the data type,index sequence of later analysis does not make the fitting of the model is significantly improved.Data to production costs,for example,conclusions of the simulated test applied in practice,empirical results show that after the continuous transformation as the qualitative variables into the model,not just failed significance tests at the variable becomes significant,you can also improve the accuracy of model fitting and forecasting.Currently this article should make such a proposition in statistics also gave no adequate theory to prove.Therefore,in the practical and theoretical aspects of research,are of great significance.From a practical point of view,the research can be applied to provide a simplified model of research ideas and methods to improve forecast accuracy.From a theoretical point of view,the study puts forward new propositions.