---

ntroduction: The use of GIS and statistical modeling to represent the possible locations of ancient sites has experienced an upward trend in recent decades (Stefan & Sîrbu, 2010 & Niknami et al., 2007 & Alirezaei et al., 2019). Currently, archaeological forecasting models are a powerful tool for preventing natural and human damage to historic-cultural resources (Danese et al. 2014) and increasing the productivity of archaeological field activities and cultural heritage management (Howard et al. al. 2016, Balla et al. 2014). Various intuitive statistical methods (qualitative and quantitative) have been used to identify ancient settlements. The most widely used method for building qualitative predictive models is the process of hierarchical analysis (Esmaili Jalvar & Heydari, 2020). The main problems with this method are the subjective judgement of experts and that intuitive models can be easily influenced by previous biases in the archaeological record. Quantitative methods can limit bias, which helps reduce the impact of this problem and can improve the reproducibility of the work for use in other settings. The most common spatial formulas for predictive modelling are: maximum entropy MaxEnt (Noviello, et al. 2018, Nicu at el, 2019),of logistic regression (Wachtel, et al. 2018), argument is evidence. (Sharafi, et al. 2016), Statistical modelling, a perspective to find suitable areas for the selection of prehistoric settlements, has been well used by geographers and archaeologists (Sharafi, et al. 2016, Verhagen & Dragut, 2012, Kaimaris, 2018). This method can be effectively considered as a form of archaeological exploration. The idea of creating a predictive model using the MaxEnt method in the eastern part of Kurdistan and evaluating such studies in archeology, as well as determining settings and making suggestions for optimizing such approaches, are among the objectives of this article. At the same time, using a statistical approach based on GIS and the use of prediction models, as well as using the data of the archeological survey of the eastern part of Kurdistan, it suggests the most favorable positions for the formation of settlements of the Iron Age. On this basis and using the environmental and archeological data of Bijar & Dehgolan cities, the prediction model for Qorveh city was presented using MaxEnt method. Due to the lack of access to the archeological data of Qorveh city, the scope of the study includes two parts, experimental and prognostic, so despite the lack of the desired data, the prediction in Qorweh region is based on the data of the experimental area (Bijar & Dehgolan) to be presented as a study method with related approaches and suggestions for other similar studies. Research method: Collecting the information of this research based on the field method, library (descriptive-analytical), using the geographical information system (GIS) preparation and interpretation of GIS maps, for analysing the settlement habitats of 96 ancient sites of the Iron Age in Eastern Kurdistan (Bijar, Qorveh, Dehgolan), and applying the prediction model of the MaxEnt model in predicting the distribution of ancient sites of the Iron Age in Eastern Kurdistan . Data: The geographical location of eastern Kurdistan includes the cities of Bijar, Qorveh and Dehgolan. According to its geographical conditions, which is the middle of the northwest, the central Zagros and the west of the central plateau, in terms of natural geography, Kurdistan province has special forms of unevenness. The Zagros Mountain range that stretches along the northwest-southeast. have formed a number of separate mountains. The ruggedness of Kurdistan, which is studied under the title of the Central Zagros Mountainous Region, includes two western and eastern parts. These two parts are different in terms of height and height and the type of land. The study area includes the political geography of the cities of Bijar, Qorveh and Dehgolan. The aim of this study is to predict the archaeological sites of the Iron Age in the city of Qorve as best as possible based on the data of the surrounding areas such as Bijar and Dehgolan. Since similar climate and landscape prevail in the eastern part of Kurdistan, Iran, the prediction of the place of origin of the sites of Qorve city is presented based on MaxEnt prediction model. In (Table 3), all the presence data (districts) of 96 districts are presented. From this number, 24 precincts (25%) were randomly selected as test data for measuring the model (for getting to know the control group or test data, see: Wachtel et al, 2018:33). And 72 areas (75%) were used for training. The MaxEnt model in this article defines two domains, the first domain, known as the education domain, includes the number of 96 presence data (areas) restricted to the Iron Age areas of Bijar and Dehgolan cities in Eastern Kurdistan (Figure 2: A). The other area, known as the prediction area, includes the city of Qorveh predicted with the above presence data, the most favourable places for the formation of Iron Age settlements. Result: The experimental domain of this modeling has 96 presence data (environment) with Iron Age chronology, and 25% of these data were introduced into the model as test data and 75% as training data. MaxEnt modeling introduces the most influential variables by examining each variable. Factors such as vegetation cover and land use, distance from the village, and distance from water sources were introduced as the most influential variables on the model results. In this case, a statistical MaxEnt analysis of the other variables is also performed. Examination of the elevation variable shows that this variable has the greatest influence on the sites in the elevation range of 1378-1400. The rivers had the greatest influence on the sites at a distance of 1000 m, at a distance of 1000 m to 3000 m they had the least influence, and at a distance of 3000 m to 5000 m they again had an effective influence on the location of the sites, and it was found that at a distance of 2000 m from the villages the chances for the formation of enclosures were higher than at further distances. The predictive map is divided into four groups based on the specific threshold mentioned earlier. Areas with very high, high, medium and low desirability. According to this classification of areas and areas with high desirability, 10.5% of the total area of the model is included, and 59% of the premises (attendance data) are located in this area. The areas of high, medium, and low desirability include 30, 6.5, and 4.5% of the sites located in these areas, respectively. Considering that the range of the predictive range is very high and low, and at the same time includes the largest percentage of the ranges, the model can be considered as a predictive application. Based on the results of Kawam, whose relationship has already been mentioned, a value close to 1 is associated with the model of the approximation value of 1, which means that the accuracy of the model has been confirmed by archaeologists. The detection value for the test area as well as the prediction area is 0.82 and 0.86, respectively. This model has shown that it is possible to perform predictive modelling for an area without presence data, and at the same time, the optimal configuration of the MaxEnt model will allow other archaeologists to study environmental variables that are influential in archaeological studies. to identify and analyse natural and human risk areas to manage cultural resources. The correct use of such models will lead to the advancement of archaeological theories in the field of cultural landscape studies an archaeological studies manage cultural resources