Purpose: This research has presented a model for evaluating the 
entrepreneurial behavior of public sector managers. This research is 
applied in terms of purpose and exploratory in terms of mixed data 
collection method.
Materials and Methods: In the qualitative phase, interviews with 10 
people were used in a non-arbitrary (targeted) way and the snowball 
method was used to achieve theoretical saturation and to screen the 
findings from 30 people using a judgment method to reach a consensus. In 
the quantitative phase of the research, 1006 managers of government 
agencies in Golestan province and the size of the statistical sample were 
estimated to be 278 using the Karjesi and Morgan table. The sampling 
method at this stage was carried out by a simple random method in such a 
way that managers were divided into three levels: top manager, middle 
manager and lower level manager and the sample was selected from among 
them.
Findings: The results of exploratory factor analysis showed that the 
structure of the management system, the existence of innovation and 
creativity space, strategic vision and effective communication were 
identified as the main dimensions and these dimensions were confirmed 
using confirmatory factor analysis. The results of the evaluation of the 
current state of the components and dimensions showed that the structure 
of the management system, the existence of innovation and creativity, 
strategic vision and effective communication are more than the average 
level and are at a suitable level. Prioritization and causal relationships 
showed that among the sub-factors of networking in the organization 
(influential), feedback systems (influential), environmental feedback 
(influential), expansion and use of interpersonal relationships (influential), 
freedom in projecting innovative ideas (effective), encouragement And 
encouraging ideas (effective), implementing new ideas (effective), 
flexibility to environmental changes (effective), taking advantage of 
opportunities (effective), vision in the organization (effective) is in the first 
to tenth priority.
Conclusion: Finally, the neural network model with the specification of 
the number of input membership function of three functions, and the 
triangular function with the output of the fixed membership function, grid 
partition regularization method and with the number of repetitions of 50 
Epochs and the optimal hybrid training method and Mamdani fuzzy system 
has a suitable system for All dimensions have been.