multivariate regression

There are many factors in the use of cars or citizens' reluctance to use a certain type of public transportation, but this paper want to budget public transportation and guide citizens to use public transportation. For this purpose, this research presents a dynamic model for the public transportation system of Mashhad using regression equations in order to achieve sustainable development. In this research, using the data related to public transportation and the opinion of experts in this field, the circular causal model of public transportation in Mashhad city was designed and simulated using Vansim software. After checking and adjusting the circular causal model, the flow diagram of the model was drawn. Then, the validity of model was formulated and tested. According to the results of the model, if the bus be free, the bus allocates about 18.06 percent of the trips, and if three thousand tomans are considered, this ratio decreases to about 11.22. It means every 1000 tomans increase in the fare rate, while keeping constant for the fare other modes, we have a 2.28% decrease in the share of bus from intra-city trips. Finally, according to the opinion of experts, 4 scenarios were presented, the results of each of which were analyzed on the development of the sustainability of the public transportation system in Mashhad.

Accurate travel time prediction is one of the important issues in the field of traffic and transportation that can significantly affect the daily life of people and organizations. In this research, four different machine learning methods including linear regression, multivariate regression, random forest and deep artificial neural network were trained to predict travel time. The purpose of this research is to predict travel time for use in intelligent traffic systems and to use and compare several new methods, including deep neural network and random forest regression, as well as considering new parameters in the computations such as weather conditions, traffic flow, travel time, and accidents and the traffic locking points compared to other studies are the innovation and comprehensiveness of this study compared to other studies. In the design and implementation of this research, real traffic data taken from Google map was used and analyzed. This data includes information such as traffic conditions, season, time of day, weather conditions, and route characteristics. The results of this research show that the deep neural network (DNN) model with R2 equal to 0.833 has a very good performance among the investigated models. This model explains 0.833% of the variance of the data and the distribution of the residuals in it is relatively central with a mean of zero and a distribution close to normal. The linear regression model with R2 equal to 0.615 has a poorer performance than DNN and explains 0.615% of the data variance. But the random regression model with R2 equal to 0.955 has one of the best performances in competition with DNN and explains 0.955% of the data variance. MSE and RMSE parameters were also used to evaluate the performance of the models, and as a result, a multidimensional comparison was made between the models, and the random forest model resulted in the lowest error values. Since in the collected traffic data, traffic accidents and consequently traffic locking points are also used in the models, and considering that the random forest model is more effectively adapted to the data despite the presence of noise and anomaly, the R2 value of this model is higher than R2 of Deep neural networks, due to the overfitting nature of Deep Learning methods.

Intersections as traffic junctions of inner-city thoroughfares are one of the most important geometric locations of pedestrian and vehicle interference, where a large number of pedestrian accidents occur every year. Therefore, the main purpose of this study is to analyze the effective factors in pedestrian accidents based on previous accident records and to investigate the impact of each factor on the number of pedestrian accidents at intersections using Poisson regression and Multivariate regression statistical analysis. Another purpose of this study is to present a GIS-based approach in order to identify and Spatial distribution of high-traffic intersections in Rasht metropolitan areas. In this study, fourteen independent factors obtained from police reports and fieldwork were divided into three general groups: human factors, weather conditions, and traffic-geometric factors. Accordingly, Poisson regression was used to analyze the direct effect of these factors on the number of pedestrian accidents at intersections, and multivariate regression was used to analyze the effect of three general categories of pedestrian crossings on pedestrian accidents at intersections. The results of the study showed that the factors of driver age 11%, rainy weather conditions 7%, and lack of traffic lights 40% had the greatest impact on increasing the number of pedestrian accidents at intersections. Also, the highest frequency of intersections with pedestrian accidents occurred in the second and fourth urban areas and the intersections of Toshiba Square, Heshmat Crossroads, and Gas Square were recognized as the most accident-prone intersections in Rasht.

Driving accidents are a serious problem in the world. Human, the road, the vehicle and the environment are the four causes of driving accidents. The most effective cause of accidents is the human factor. Therefore, in the field of traffic safety, social science studies, especially psychology studies, are necessary. Any attempt to get to know the personality and improve driving behavior can save many lives. Personality tests are used to identify a person's personality. The research instrument was a questionnaire consisting of three sections: demographic questions, high risk driving behavior questionnaire and Myers-Briggs personality test. Traffic high-risk behaviors included 38 items that were classified into four factors: aggressive behavior, fatigue and distraction, lack of agreeableness and lack of attention to their health. And personality preferences obtained from the Myers-Briggs test include (Extrovert-Introvert, sensor-intuitive, Thinker- Feeler and Perceiver -judger). Multivariate regression was used to measure the predictive power of the number of accidents. The results of regression analysis showed that among the personality dimensions of Myers-Briggs test, the perceptual variable has the most significant predictive power of accident variable. Also, among the high risk driving behaviors, aggressive behavior and distraction have the most significant predictive power of accident variables.

In today 's world, optimizing a successful business depends on using all the resources that make it superior to its competitors. Location-based marketing or Geomarketing leads to critical and effective decisions by analyzing different geographical areas. Spatial information systems marketing is more agile in strategic decision making. In this age where data dynamics are so important, the use of spatial data infrastructure (SDI) can create a platform for spatial data sharing. Spatial Data Infrastructure (SDI) with instant sharing of spatial data can provide a dynamic platform. SDI-based Geomarketing fixes the flaws and shortcomings of spatial information layers in GIS-based Geomarketing. The main advantage of this model compared to previous models, in addition to information dynamics, is that there is no need for an operator to record and store information and produce layers of location-based information in alternating time periods. This is an applied research in terms of purpose and is based on a descriptive method that includes a set of methods that aim to describe the conditions or phenomena under study. In terms of implementation, part of this research is collected in the form of libraries and documents using the theoretical foundations and background of previous research, and the other part is done experimentally and by collecting information from the base statistical reference authorities.Accordingly, in this research, a new model of location-based marketing is presented, which uses spatial data infrastructure for the first time. In this article, we seek to answer the questions of whether the use of Geomarketing based on spatial data infrastructure has an advantage over GIS-based location-based marketing? Is it possible to prioritize the optimal areas by sharing important indicators from different databases of executive agencies in the field of marketing of Semnan banks? In this regard, using this model and based on the data available in 4 databases of related executive agencies, the city of Semnan is divided into 139 urban areas or statistical areas. Afterwards, using the geoportal infrastructure of Semnan province spatial data located in the Management and Planning Organization of Semnan province, the desired registration information layers were shared and model’s maps were extracted. Subsequently, by examining 150 demographic and economic indicators and examining their correlation coefficient with the number of bank branches, it was found that the indicators of literacy rate, household size, population density, distance from the city center, number of important businesses, income decile, number of apartments and number of schools are most relevant with the number of bank branches in each region. Then a model was estimated using multivariate regression. In this model after estimating the model coefficients, the number of businesses index with a coefficient of 0.598 has the greatest impact on the number of bank branches. According to the results of this model, area No. 62 of Semnan city has the most favorable conditions in terms of banking marketing indicators. So the main advantage of this model compared to previous models, is that there is no need for an operator to record and store information and produce layers of location-based information in alternating time periods in addition to information dynamics. In this model, a dynamic model can be achieved by using dynamic information by sharing layers of spatial information in the context of spatial data infrastructure, in addition to maintaining the intellectual property of information. This research is supported by the GIS unit of the Management and Planning Organization of Semnan Province in Iran.

Many engineering structures have been built on the sandstones. The main purpose of this study is to estimate the uniaxial compressive strength (UCS) and modulus of elasticity (Es) of sandstones using regression models. For this purpose, petrographic studies, compressional wave velocity (Vp), porosity, density and uniaxial compressive strength tests were performed on dry and saturated samples of sandstones prepared from Mosha village in the northwest of Damavand city. The studied sandstones were classified as feldspathic litharenite and litharenite. Due to the effect of moisture on the physical and mechanical properties of these sandstones, the density and Vp of the samples in the saturated state compared to the dry state have increased by 4 and 20%, respectively. In contrast, UCS and Es have increased by 18% and 25%, respectively. The results of simple regression showed that the most accurate relationship (the highest correlation coefficient and the lowest error) of porosity, Vp and density with UCS and Es are logarithmic, linear and quadratic polynomials, respectively. Based on the determination coefficient (R2=0.5-0.77) and the errors (RMSE=10.29-18.26; MAPE=1.70-2.80), the relationships presented by simple regression method for estimating UCS and Es showed high accuracy. The Vp and porosity also have the greatest impact on UCS and Es. Examination of residual variance homogeneity graphs at the predicted value levels, determination coefficient and error of the methods showed that multivariate regression (R2=0.73-0.74, RMSE=13.36-13.56, MAPE=1.06-1.22, Durbin-Watson= 1.56-1.70) has a high accuracy for estimating UCS and Es as compared to the simple regression.

Deformation Modulus of Rock Masses is one of the most important rock mass parameters that is used to measure the displacement of rocks under loading. Dynamic loads such as earthquakes and explosions are also important in rock engineering, as are static loads. Therefore, it is necessary to consider dynamic parameters as much as static properties. Dynamic deformation Modulus of Rock Masses is determined using Tomography and seismic refraction tests, which are time-consuming and cost-effective. Therefore, in this study, indirect methods based on the combinations of engineering classification systems and laboratory results were used to estimate the dynamic deformation modulus of rock masses. Due to the easy and comprehensive measurement of the Geological Strength Index (GSI) classification system and ultrasonic tests in the laboratory, the data of the five large dams in Iran were gathered and analyzed. The multivariate regression analysis showed that the dynamic deformation modulus of the rock masses can be estimated by an acceptable approximation with combination of GSI and the ultrasonic tests results.