Development of land value algorithm for establishing an effective cadastral system in Erbil City

Land value is one of the economic issues of cadastral systems which is the base of sustainable urban and regional planning. The current paper concerns the estimation of the land values according to many essential factors, which are adopted as ten variables generally. Among these ten parameters, the frontage of the parcel (width), the value of rent, the width of streets, and the level of services represent the most effective parameters that play the main role in process of land price estimation over the Erbil City. The current research introduces the nature of land values and their homogeneous distribution and evaluates the suggested algorithm of land price estimation as one of the efficient factors that affect the national economic situation. The data collection was done for 100 parcels in different locations within the Erbil city boundary, which is being selected to apply the linear multiple regression equation to find the coefficients of the effective factors and to define the correlation between them. The obtained results of the linear multiple regression equation show that the level of existing services and the value of the rent have the maximum effect among these four factors, and they have the maximum correlation with the land price, whereas the road’s width has the minimum correlation among them. The worked-out algorithm for land price estimation (which is a vital issue of the modern cadastral systems), is recommended to be applied by the institutions and organizations concerning the land prices and land taxes task.


Introduction
The cadastral issues represent an active field of the economic sector and strategic planning. It is the basis for legal aspects like ownership as well as fiscal aspects like taxation duty of urban territories. The efficient cadastral system provides the necessary data and the required information for land use planning and infrastructure construction. The modern cadastral systems were developed to meet the requirements of the parcel's ownership and mortgage rights and hence the taxation issues. Today a cadaster is also used as a basis for planning assignments like the dedication of land [1]. Land value is one of the three basic components of the cadastral system with land use and land tenure [2].

Case study (Erbil City)
Erbil city is the fourth major city of Iraq after Baghdad, Mosul, and Basra. lies at the longitude 43.9930° E and the latitude 36.1901° N. To define the most vital factors that affect land value, 120 questionnaire forms are prepared and distributed to cover the different zones of Erbil city. One hundred of them are accepted and the other twenty are neglected due to incompletion or inaccurate information [3]. Figure 1 shows the questionnaire distribution areas. Eighty questionaries' forms have been used to establish the required mathematical model and to find the desired coefficients. The questionnaire forms are distributed to the real estate offices and the relevant companies in different zones of Erbil city to cover all the urban territories within the city's administrative boundaries. Twenty of the questionnaires' forms are used for assessing the obtained results, which is done according to ten adopted parameters as follows: 1-Parcel's width. 2-Width of the existing streets. 3-Size of parcel area. 4-Location of the parcel according to the Central Business District (CBD). 5-The value of the rent. 6-Parcel's orientation. 7-Existence of the opposite green area. 8-Parcel's position concerning neighbor parcels (corner). 9-The potential of commercial exploitation. 10-The level of existing services. However, the research's methodology can be summarized in the flowchart below (see Figure 2).
It is worth mentioning that the results of linear multiple regression show that six of the 10 parameters above have a weak correlation with the land price (which is 0.2 -0.3), therefore they should be regarded as a not significant parameter that can be neglected. Thus, in the next development, just four parameters will be considered as the variables that significantly affect the estimation of the parcel's price (see Equation 1).

Linear multiple regression equation
The collected data are converted into a numerical form as a preparation for statistical analysis and then to establish the mathematical model according to the following multiple regression algorithm. Where: Pi -is the predicted land price of m2, which represents the dependent variable in the equation. Pₒ -is the essential price of m2 (basic price) in the ith zone. B₁, B₂, B₃, & B₄ -are the coefficients of the four effective parameters (variables) which represent the slope of the regression line. X₁ -is the parcel's width (frontage of parcel). X2 -is the width of streets in the ith zone. X₃ -is the value of rent in the ith zone. X4 -is the level of services in the ith zone. n -is the number of the selected parcels in all zones, which represents the number of equations in the mathematical model [5]. The size of the established mathematical model is eighty simultaneous equations, for the four effective parameters. Thus, the mathematical model (system in Equation 1) can be rewritten as follow: The software called Statistical Package for Social Science (SPSS) version 28 (see figure 3), has been used for the multiple linear regression application and the statistical analysis process, and then for determining the degree of dependency between the variables and the predicted land price [4]. The results of the application of the linear multiple regression and the obtained correlations between land price and each one of the four effective parameters are illustrated in tables (1 & 2) below. The data in Table 1 illustrates the obtained coefficients of the four parameters: parcel's frontage (width), street's width, the value of rent, and level of services. Meanwhile, the data in Table 2 show the degrees of dependence between land price and the corresponding four parameters through the obtained values of the Pearson correlation between them [6]. From the table the B is the constant value in linear regression and Beta is the linear equation with a out constant value in land value must include the B in the equation to achieve the actual land and price as reality , The is to test to find out if small m n is significant with the population mean ? because the computed value is more than the table value which is 2 t test is significant , and the significant level sig. for all independent variables are less than 5% that's mean 95% there is relation between independent and independent variables [7].
Then, the price of m2 can be estimated in each zone separately according to the variables, parcel's width, street's width, the value of rent, and the level of existing services as in Eq. 3 above.
Notice, that the rent's value plays the most effective role in process of price estimation since it has a greater correlation with the price, whereas the parcel's frontage (width) has less effect on the predicted price since it has the smallest correlation among all the four parameters [8].

Evaluation of the obtained algorithm
In the process of evaluation of the obtained land value algorithm, a questionary of 20 selected parcels has been prepared for this purpose. The predicted prices for these 20 parcels are compared with the actual prices obtained by the real estate offices in different zones of the case study. Table 3 shows the results of the evaluation process and the differences between the computed price and the up-to-date actual price of one m2 in different zones over Erbil city territories [9].
These differences (deviations of the price) are used to find the relative price deviation (RPD), which is the (deviation/price) ratio.
The obtained results of the evaluation (see table 3) indicate that the price deviations of the tested parcels have a reasonable value with a negative sign in most zones. The analysis of the values of relative price deviation, which does not exceed 0.084 and its average value equal to 0.039, shows the precision of the computed prices and consequently the accuracy of the used algorithm [10]. The histogram of the obtained regression standard residuals is shown in Figure 4 below. The parcel width and street width are uploaded from the digital map (Pimpler, 2017), and the rent with service level is assumed according to governmental regulations in that zone [11]. The linear regression equation (Equation. 3) is applied to calculate the price of one m2 of each sample and then the price of the samples using the predicted price of m 2 multiplied by the area [12]. Furthermore, the taxes of the selected samples are calculated based on governmental tax regulations, which presented the tax's ratio as equal to 0.01 multiplied by the predicted price. The layer of taxes for all samples is illustrated in Figure 8 below [13].  Spatially can apply this formula in ArcGis Environment by Field calculator and in property y of attribute table Generating the land price and parcel tax for all parcels in the City of Erbil by in putting the four variables in the formula.
The analysis and results find that the center of the city does not affect forever the land value although it remail as an expensive area, but the other nodes and districts may compete with the center and overcome it the price of land. The assessors and arousers must look at the whole city in terms of price and not on the neighborhoods of the plot to be valued only to obtain a fair, comprehensive, and accurate estimation.

Conclusions
1. The predicted land price can be depending on ten parameters in general but, only four of them (Parcel's frontage, Street width, rent value, and level of services) are practically affect the land value since they have a maximum linear correlation with the land price. 2. The other six parameters (parcel orientation, distance from the city canter, closing to a green area, corner, commercial area, and location) have a minimum correlation with land price and therefore they are regarded as not effective variables that can affect the predicted land value. 3. The linear multiple regression equation gives an accurate estimation of the land price if it is applied with the real variables and efficient parameters. 4. The applied linear regression algorithm based on the four effective parameters can be applied in Erbil city territories since it gives fair and accurate predicted prices overall case study area. 5. The evaluation process of the worked-out algorithm shows that the average deviation of the predicted prices is about 3.9%, which can be understood as the average of the algorithm's precision. 6. It is recommended to use the obtained algorithm (Eq. 3) over the Erbil urban territories for estimation of the land price as one of the sources desired for the land exploitation, taxation duties, and other economic activities of the cadastral systems.

Declaration of competing interest
The authors declare that they have no known financial or non-financial competing interests in any material discussed in this paper.

Funding information
No funding was received from any financial organization to conduct this research.