The Computational Model of Political Trust in the Islamic Republic of Iran with the Help of Data Mining

Measuring political trust is necessary to establish the stability of the society in the events. This research tries to obtain the calculation model of political trust with the help of big data of 1649 variables in different fields, resulting from the activism of Iranian citizens based on "observational big data". The research method is a combination of quantitative and qualitative methods (with the dominance of the quantitative method); Therefore, first, the concept, dimensions and variables of political trust were investigated based on the document analysis method and with the phishing tool. Then, according to the construct theory of the concept of political trust, the related variables were extracted with the help of interviews with experts, and then the KDD method was used to analyze the big data. Also, with the help of machine learning, two supervised and unsupervised data mining methods were analyzed and the best algorithm was selected according to the two criteria of lower risk and greater utility. With the conducted investigations, the best model of supervised learning under classification and Neural Net algorithm was introduced. Then, with the IBM SPSS Modeler tool, the data was "classified" in four steps, screening, discovery of correlated data, normalization and modeling. It should be noted that in the modeling stage, the influence coefficients of 13 variables obtained from the previous stages were extracted and their ratio in the neural network was shown according to the effect of the middle layers (hidden layers) on the political trust based on the neural network.