جستجوی مقالات مرتبط با کلیدواژه « Multivariate model » در نشریات گروه « زمین شناسی »

Combination of Adoptive Network based Fuzzy Inference System (ANFIS) and subtractive clustering (SC) has been used for estimation of deformation modulus (Em) and rock mass strength (UCSm) considering depth of measurement. To do this, learning of the ANFIS based subtractive clustering (ANFISBSC) was performed firstly on 125 measurements of 9 variables such as rock mass strength (UCSm), deformation modulus (Em), depth, spacing, persistence, aperture, intact rock strength (UCSi), geomechanical rating (RMR) and elastic modulus (Ei). Then, at second phase, testing the trained ANFISBSC structure has been perfomed on 40 data measurements. Therefore, predictive rock mass models have been developed for 2-6 variables where model complexity influences the estimation accuracy. Results of multivariate simulation of rock mass for estimating UCSm and Em have shown that accuracy of the ANFISBSC method increases coincident with development of model from 2 variables to 6 variables. According to the results, 3-variable model of ANFISBSC method has general estimation of both UCSm and Em corresponding with 20% to 30% error while the results of multivariate analysis are successfully improved by 6-variable model with error of less than 3%. Also, dip of the fitted line on data point of measured and estimated UCSm and Em for 6-variable model approaches about 1 respect to 0.94 for 3- variable model. Therefore, it can be concluded that 6-variable model of ANFISBSC gives reasonable prediction of UCSm and Em.

Tropospheric ozone is one of the main causes of respiratory problems and it hurts vegetations. In this research, statistical models based on a wide variety of regression models are presented in order to evaluate the surface ozone concentrations in hourly and daily scales in Isfahan using meteorological variables and pollutant gases as predictors. Although none of meteorological variables and pollutant gas levels has the ability to interpret the measured ozone variations in Isfahan, the results have shown there is a significant correlation between them and the ozone variations. Calculating a nonlinear bivariate model can show the general ozone fluctuations, but because of irregular fluctuations in hourly data, it can not be a proper predictor. Most of the models assigned the biggest influence to the air temperature and humidity in surface ozone production and declared that the mean surface pressure do not have an important role in the point analysis. Also increasing the oxide compositions of nitrogen increases the ozone production. In a daily scale, carbon monoxide and temperature have presented the best interpretation for the ozone concentration. The aim of this research was to present a consistent evaluation of the surface ozone using statistical methods. At beginning, the society of ozone samples, pollutant gases and corresponding meteorological data was assessed and the correlation between the ozone level and each of them or a group of them was tested, step by step. Most of data did not obey a normal curve, so in different stages, some operations were necessary to make the data closer to the normal situation. In this paper, the data from meteorological and pollution observations were used as predictors. The station was located in 32.62N, 51.66E with the elevation of 1550 m. The data consisted of: a- The data from the pollution stations: surface ozone, CO, SO2, NO, NOx, NO2 b- Meteorological data: air temperature, relative humidity, wind speed, solar radiation, air pressure. Reviewing the time series of the ozone data (24 hours) showed that there was a daily sinusoidal cycle in the ozone concentration and a sinusoidal model can easily calculate the ozone amount as a function of the hours in a day. Although a sinusoidal curve was well fitted to the daily curve of the ozone concentration, random fluctuations in the daily average were seen. These irregularities caused difficulties in presenting a single proper model to show the daily cycle of the ozone concentrations. In the next stage, an equation was gained by modulation of the daily and hourly equations to show the ensemble daily and hourly cycle of the ozone concentrations. Analysis of the results of regression models shows that between the three equation, best equation be gained from step wise method. Then, by using a backward method, 13 equation be gained. All of these equations show that the daily scale can not justify the surface ozone variations. This can be because of the act of other unknown variables or because of the nonlinear nature of the correlations between ozone levels and the predictors. However the data were preprocessed to get closer to a normal distribution. For this purpose, both logarithmic and squared forms of the data were also used eventhough they could not make a considerable change in order to transform the data to normal distributions. All of them were used beside the natural data to form more regression models. It should be noted that the nature of these kinds of data, that needs complicated process to be created, makes the correlations coefficient less strong. The resulted equations in this paper showed that the current operations could not normalize the distributions of the data. The existence of a nonlinear correlation between the ozone levels and the studied variables can be a reason for the weakness of these models. In the previous studies, the highest determination coefficient was 0.36 (Alexandrof, 2005). In this paper, the best equation nearly showed the same amounts (r = 0.304). In the backwards method, a higher coefficient was gained (r = 0.592) but because of the length and size of the equation, it is not usable. Although the regression models and the principal component analysis showed that they had a strong ability to interpretat the surface ozone fluctuations and predict its concentration, the number of their independent variables prevented them from being useful enough from an application viewpoint.