A Study on Regression Model Using Response Surface Methodology

Response Surface Methodology (RSM) mostly employs statistical regression method as it is practical, economical and relatively easy to use. The first and second order polynomial equation was developed using RSM. This polynomial model usually refers as a regression model. In this research, the objective is to find the best response surface method to model three factors and three levels parameters in machining. From the study, the Box-Behnken Design can develop a good regression model rather than Central Composite Design or Full Factorial Design. While, the second order regression model has proved to be more effective in predicting the performance of the given data set.

Research on Regional Dividing Technology for Liquefaction-Induced Lateral Spreads Based on MAPGIS

A simplified empirical statistical regression method for liquefaction-induced lateral spreads is stated in this paper. It is used to determine the corresponding calculation parameters for Tangshan city and its south region where 36 borehole data were collected. Using two multiple linear regression formulas, the free face model and the ground slope model are developed to calculate the lateral spread at each filed survey point in the region. The liquefaction-induced lateral spreads map was divided into hazard zones according to the calculated horizontal displacement at each borehole based on MAPGIS.

The Application of Statistical Regression Method on the Seismic Building Economic Loss Evaluation of Yunnan Province

More than 30 earthquake loss assessment reports in Yunnan Province in 1993 – 2003 are collected. And the building loss in each intensity recalculated based on the all building area, destruction ratio, loss ratio and replacement cost which are listed in the reports. The relation curves of building loss versus affected population, affected area in each intensity are studied respectively using statistical regression method. After regression significance test, it is found that the relation of building loss and affected population is high; however, there is no obvious relevance of building loss versus affected area. So in this paper, the quantitative relation between the housing loss and earthquake-affected population in each intensity area is established in Yunnan Taking four earthquakes during 2004 ~ 2005 in Yunnan as examples; we estimated their earthquake building economic loss and compared them with the values taken from field investigation. The result shows this regression equations are useful and applicable for evaluating the building loss caused by earthquake.

Research on Ossicular Chain Mechanics Model

On account of the complex structure of the middle ear and that it is more difficult to carry out experiments to test the nature of the mechanical components, the relevant data is hard to obtain, which has become an obstacle restricting analysis of the middle ear mechanic. Based on spatial structure and mechanical properties of ossicular chain, the paper has established functional relationship between load and members displacement with elastic principles and variation principles, in order that experimental results will be reflected in mechanical model. In the process of solving equations, we use the experimental data of a known special point or the various components function of statistical regression method and then combine them with the time shift function, so that the analytical solution of the various components will be achieved. The correctness of equation derived in this paper is verified by comparing the experimental data. So the model has provided a convenient way to obtain date in the future research analysis.