جستجوی مقالات مرتبط با کلیدواژه « الگوریتم های یادگیری ماشین » در نشریات گروه « عمران »

Road construction and maintenance organizations usually use certain criteria to qualify asphalt mixtures before use in construction. One of the most important features that is measured in the asphalt mixing and quality control plan is the Marshall asphalt stability. This study examines the use of machine learning techniques to predict Marshall asphalt stability. Due to the time-consuming and costly process of asphalt production and quality control, it is necessary to use new methods in this process. In this research, two supervised support vector machine and random forest algorithms, which are machine learning algorithms, were used to predict the marshal asphalt stability. For this purpose, the test results of 2000 asphalt samples including Granulation of aggregate, Fracture percentage, bitumen adsorption, bitumen specific gravity, actual specific gravity of materials, bitumen consumption percentage, dust to effective binder ratio and Marshall asphalt stability for training and evaluation Models were used. After modeling and evaluation, the value of R2 is 87.5 for the support vector machine method and 82.69 for the random forest. Also, MAPE, RMES and SDE values for SVM were 3.1073, 40.042 and 0.0208, respectively, and for RF were 3.1641, 41.870 and 0.0211, respectively. The results show the efficiency of the models used against laboratory methods for predicting marshal asphalt stability, which SVM method has a better performance than RF. Machine learning methods can be used to predict other parameters of the asphalt mixing plan and reduce the time, cost and human error of the tests.

Definition of the efficient ore-forming processes which are considered as mineralization controls is a fundamental stage in mineral prospectivity mapping. In this contribution, five targeting criteria of geochemical, geophysical, geological, structural and hydrothemal alteration data related to epithermal and Carlin-type Au deposits in Takhte-soleyman district, NW Iran, were integrated. For creation of multi-element geochemical layer, principal component analysis was firstly conducted on stream sediment data of 8 selected elements and it was found that PC2 is the representative of Au-As-Sb elemental association in the study area. Then, 10 fuzzified efficient evidence layers were selected based on area under the curve (AUC) of success-rate curves and prepared for generation of predictive models of Au mineralization. For this purpose, two supervised machine learning algorithms, namely multi-layer perceptron (MLP) neural network and support vector machine (SVM) with RBF kernel were used. Comparison of the generated models demonstrates that the latter is more succeeded in delineating exploration targets than the former one.