Land price estimation system

Abstract

The present invention relates to a system (1) for estimating prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets.

Description

[0001] LAND PRICE ESTIMATION SYSTEM

[0002] Technical Field

[0003] The present invention relates to a system for estimating prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets.

[0004] Background of the Invention

[0005] In today's changing economic conditions, pricing has become very important. One of the fields affected by fluctuations due to these economic conditions is the real estate sector. While conventional methods generally make estimates based on the prices of comparable real estates, the accuracy of these methods depends on many external factors. In particular, errors in the selection of comparables and the presence of outliers reduce the reliability of estimates.

[0006] For this reason, it is understood that there is a need for a system which enables more accurate and faster price estimation by using advanced artificial intelligence algorithms.

[0007] The Turkish patent document no. TR2022 / 011790, an application included in the state of the art, relates to an automatic house price calculation system and algorithm which can reveal the maximum and minimum prices of houses during the year by artificial intelligence integration by being integrated into government institutions, real estate institutions, public housing companies and websites, and the feature of which is that it comprises at least one electronic device configured to enable the user to communicate with the owners and display prices for the house that the user wants to rent; a search module that runs on the electronic device and filters according to the commands related to the location information, number of rooms, square meters and environmental conditions of the house input by the user; a database that contains rental information according to location information, number of rooms, square meters and environmental conditions for the past years; and an artificial intelligence algorithm that enables the system to be trained within the information in the database and estimates the maximum and minimum price information that may be available for the relevant house within the year.

[0008] Summary of the Invention

[0009] An object of the present invention is to realize a system developed with the aim of enabling reliable and fast price estimation in the real estate sector.

[0010] Another object of the present invention is to realize a system developed with the aim of estimating prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets.

[0011] Detailed Description of the Invention

[0012] “Land Price Estimation System” realized to fulfil the objectives of the present invention is shown in the figure attached, in which:

[0013] Figure l is a schematic view of the inventive system.

[0014] The components illustrated in the figure are individually numbered, where the numbers refer to the following:

[0015] 1. System 2. Database

[0016] 3. Server

[0017] The inventive system (1) developed with the aim of estimating prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets comprises; at least one database (2) which is configured to keep a record of data in the form of sales prices, rental prices or other relevant financial information of comparable real estates; at least one server (3) which is configured to identify other real estates (comparables) similar to the real estate to be valued by using artificial intelligence algorithms; to collect data in the form of sales prices, rental prices or other relevant financial information of comparable real estates from real estate records, real estate agents, listings or public data; to adjust the selling prices of comparable real estates according to the characteristics of the evaluated real estate; to calculate the average price of the comparables after the adjustments are made and to accept this value as the estimated market value of the evaluated real estate; and to create a detailed report in the form of a list of the comparables used, the adjustments made and the final value estimate by collecting all the analyses and calculations made in a report.

[0018] The database (2) included in the inventive system (1) is configured to establish communication and to exchange data with the server (3) by using any communication protocol. The database (2) is configured to keep a record of the data received by the server (3) from various data sources.

[0019] The server (3) included in the inventive system (1) is configured to establish communication and to exchange data with the database (2) by using any communication protocol. The server (3) is configured to use machine learning models. The server (3) is configured to perform outlier elimination with the Tukey method. The server (3) is configured to determine the upper and lower limits and to consider data outside these limits as outliers using the interquartile range (IQR) of the data with the Tukey method. The server (3) is configured to determine IQR limits as above 1.5 IQR and below 1 IQR in the Tukey method. The server (3) is configured to use a hierarchical structure at the neighbourhood, district and city level in the outlier elimination process through the unit price of residential zoned land. The server (3) is configured to enable examples to be extracted from the dataset at the neighbourhood level. The server (3) is configured to enable analysis to be performed at the district level if there are not enough examples at the neighbourhood level, and at the province level if there are not enough examples there either. The server (3) is configured to enable outliers to be detected by maintaining the homogeneity of the data. The server (3) is configured to enable data to be received from various sources. The server (3) is configured to enable the quarter- and location-based price data to be grouped into latitude- and longitude-based listings and the average price values to be given to the model as a variable on the basis of quarters of the year. The server (3) is configured to enable the distance data to the centre of the region to be given to the model as a variable to be used in training by calculating the distance to the centre of the region based on the latitude and longitude values of the lands. The server (3) is configured to enable neighbourhood-based density data to be given to the model. The server (3) is configured to enable region category and title deed type data to be given to the model in the form of village, town, district and detached title deed, share title deed. The server (3) is configured to enable Floor Area Ratio (FAR) data used to determine the total building area that can be built on a land to be given to the model. The server (3) is configured to enable the economic variable data to be given to the model. The server (3) is configured to enable the inflation rate (CPI) value realised one month before the date of the relevant listing to be given as a variable to the model on a listing basis as economic variables. The server (3) is configured to enable the target variable to be calculated with the dollar exchange rate by taking into account the dollar exchange rate specific to the listing date if it is the land unit price, and the model to learn the land price in dollar basis. The server (3) is configured to further normalise the distribution of the target variable by taking the logarithm of the target variable, and thus to reduce the impact on the model of outliers not eliminated by outlier elimination. The server (3) is configured to comprise the LightGBM (Light Gradient Boosting Machine) machine learning algorithm that operates quickly and with high performance on large data sets. The server (3) is configured to provide fast training and estimation time on large data sets with the LightGBM algorithm. The server (3) is configured to enable large amounts of data to be processed quickly and efficiently with the ability of the LightGBM algorithm to scale independently of the size of the data set. The server (3) is configured to enable complex data relationships to be learnt and highly accurate predictions to be made with the LightGBM algorithm in land price estimations. The server (3) is configured to enable the transparency of the model to be increased and the key factors affecting land prices to be understood by facilitating the determination of which features are more important in the decision-making process of the model with the LightGBM algorithm. The server (3) is configured to enable system resources to be used efficiently even when working with large data sets by means of the optimised memory usage of the LightGBM algorithm and advantages to be provided in environments with limited hardware resources. The server (3) is configured to enable situations where land prices usually vary over a wide range to be kept under control by means of the properties of the LightGBM algorithm to be robust to outliers and to minimise the negative impact of these values on the performance of the model.

[0020] Industrial Application of the Invention

[0021] By means of the inventive system (1), it enabled to estimate prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets. Within these basic concepts; it is possible to develop various embodiments of the inventive “Land Price Estimation System (1)”; the invention cannot be limited to examples disclosed herein and it is essentially according to claims.

Claims

CLAIMS1. A system (1) for estimating prices by analysing historical sales data, regional economic indicators and real estate characteristics in the real estate sector with a powerful artificial intelligence model that can process and analyse large data sets comprising; at least one database (2) which is configured to keep a record of data in the form of sales prices, rental prices or other relevant financial information of comparable real estates; at least one server (3) which is configured to identify other real estates (comparables) similar to the real estate to be valued by using artificial intelligence algorithms; to collect data in the form of sales prices, rental prices or other relevant financial information of comparable real estates from real estate records, real estate agents, listings or public data; to adjust the selling prices of comparable real estates according to the characteristics of the evaluated real estate; to calculate the average price of the comparables after the adjustments are made and to accept this value as the estimated market value of the evaluated real estate; and to create a detailed report in the form of a list of the comparables used, the adjustments made and the final value estimate by collecting all the analyses and calculations made in a report.

2. A system (1) according to Claim 1; characterized by the database (2) which is configured to establish communication and to exchange data with the server (3) by using any communication protocol.

3. A system (1) according to Claim 1 or 2; characterized by the database (2) which is configured to keep a record of the data received by the server (3) from various data sources.

4. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to establish communication and to exchange data with the database (2) by using any communication protocol.

5. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to use machine learning models.

6. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to perform outlier elimination with the Tukey method.

7. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to determine the upper and lower limits and to consider data outside these limits as outliers using the interquartile range (IQR) of the data with the Tukey method.

8. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to determine IQR limits as above 1.5 IQR and below 1 IQR in the Tukey method.

9. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to use a hierarchical structure at the neighbourhood, district and city level in the outlier elimination process through the unit price of residential zoned land.

10. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable examples to be extracted from the dataset at the neighbourhood level.

11. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable analysis to be performed at the districtlevel if there are not enough examples at the neighbourhood level, and at the province level if there are not enough examples there either.

12. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable outliers to be detected by maintaining the homogeneity of the data.

13. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable data to be received from various sources.

14. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the quarter- and location-based price data to be grouped into latitude- and longitude-based listings and the average price values to be given to the model as a variable on the basis of quarters of the year.

15. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the distance data to the centre of the region to be given to the model as a variable to be used in training by calculating the distance to the centre of the region based on the latitude and longitude values of the lands.

16. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable neighbourhood-based density data to be given to the model.

17. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable region category and title deed type data to be given to the model in the form of village, town, district and detached title deed, share title deed.

18. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable Floor Area Ratio (FAR) data used to determine the total building area that can be built on a land to be given to the model.

19. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the economic variable data to be given to the model.

20. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the inflation rate (CPI) value realised one month before the date of the relevant listing to be given as a variable to the model on a listing basis as economic variables.

21. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the target variable to be calculated with the dollar exchange rate by taking into account the dollar exchange rate specific to the listing date if it is the land unit price, and the model to learn the land price in dollar basis.

22. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to further normalise the distribution of the target variable by taking the logarithm of the target variable, and thus to reduce the impact on the model of outliers not eliminated by outlier elimination.

23. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to comprise the LightGBM (Light Gradient Boosting Machine) machine learning algorithm that operates quickly and with high performance on large data sets.

24. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to provide fast training and estimation time on large data sets with the LightGBM algorithm.

25. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable large amounts of data to be processed quickly and efficiently with the ability of the LightGBM algorithm to scale independently of the size of the data set.

26. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable complex data relationships to be learnt and highly accurate predictions to be made with the LightGBM algorithm in land price estimations.

27. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable the transparency of the model to be increased and the key factors affecting land prices to be understood by facilitating the determination of which features are more important in the deci si on -making process of the model with the LightGBM algorithm.

28. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable system resources to be used efficiently even when working with large data sets by means of the optimised memory usage of the LightGBM algorithm and advantages to be provided in environments with limited hardware resources.

29. A system (1) according to any one of the preceding claims; characterized by the server (3) which is configured to enable situations where land prices usually vary over a wide range to be kept under control by means of the properties of the LightGBM algorithm to be robust to outliers and to minimise the negative impact of these values on the performance of the model.

Images