Kerentseva Marina

Introduction

The booming real estate market led to a sharp increase in the number of companies providing real estate services, which exacerbated the competition and more stringent requirements on the level, speed, convenience and quality customer service.

There are two main problems arising in the process of real estate agency:

1. Search and selection of variants, corresponding to a given set of parameters and criteria. This situation arises when the agency concludes with the buyer an agreement on cooperation in order to find the necessary funding variant.
2. Real estate appraisal. This problem must be addressed whenever seller appeals an agency to set a real estate variant for sale.

The efficiency of the whole enterprise depends on the quality of the first phase, the success of which requires multiple processing of the array of actual proposals of the market, and when evaluating real estate – also analyze the current market conditions, prevailing market segmentation and steady prices in each segment.

The substantial increase in demands for speed and quality of performance of these steps requires the development and implementation of the system, providing the possibility of analyzing information about the current state of the property market and allows to make an assessment of property values.

Thus, the relevance of the work is determined by the need to develop mathematical and algorithmic models of the functioning of an expert system for determining the price of real estate, as well as the development of a software architecture suitable for the practical implementation of the system.

The purpose and objectives of the research

The aim of the research is generalized mathematical models and principles of operation of computer automation systems realtor company, which is one of the links in the process of making marketing solutions company.

To achieve this goal need to address the following main tasks:

• Analysis of data processing, evaluation of real estate;
• Development of search suggestions on the real estate market that match vague terms;
• Development of a mathematical model of evaluation of properties based on the method of sales and database content.

Topicality

A significant increase in speed and quality of all these actions requires the development and implementation of expert system (ES), providing the possibility of analyzing information about the current state of the property market.

Thus, the relevance of the work is determined by the need to develop mathematical and algorithmic models of the functioning of the ES realtor companies, as well as the development of a software architecture suitable for the practical implementation of the system.

Alleged scientific novelty

The scientific novelty of this work is to develop an expert system based on fuzzy rules, which allows perform real estate appraisal. Appliances Fuzzy Logic will also implement fuzzy query to the database.

Expected practical results

The results of this work is to construct rules of expert system for determining the value of the property, given a detailed description of each object, a software implementation designed tools. Will be implemented object-oriented model of software. This assumes a partition of the system into several subsystems: a subsystem of communication with the user, a subsystem of knowledge acquisition, storage and analytical data base of rules, machine inference, subsystem analysis and reporting.

Problem statement

After analyzing the list of tasks can come to the conclusion that the solution lies in building an expert system that can partially replace the specialist expert in solving the problematic situation, while allowing decision-making under conditions of incomplete and / or fuzzy information [5].

In the domain of sale of real estate are of key importance not only precise, mathematical-based data, and models containing qualitative information, which includes long-term operating experience and important information about this area of expertise. Language of fuzzy sets and algorithms are currently the most appropriate mathematical tool that allows to minimize the transition from verbal qualitative description of the object to a numerical quantitative estimate of its condition and to formulate on this basis, simple and efficient algorithms, that is it possible to simulate the human thinking and human problem solving ability [5].

Building a base of rules and results of research.

The initial task is to build a rules expert system for determining the value of the property, given a detailed description of each object. Base rules built an expert on the set of real data Voroshilov, Kiev, Lenin and Kirov districts for the year 2009 in one of the agencies of Donetsk. To meet the goal these steps are followed:

1) Entered the input linguistic variables to the basic term-sets:
Location = (notPrestigious, middle, prestigious, veryPrestigious);
RoomNumber = (one, several, many);
State = (unsatisfactory, satisfactory, good, excellent, euro);
Houseroom = (small-size, average, large-size);
Distance-to-transport = (near, average, distant);
Output linguistic variable: Price = (low, average, high, very-high)

2) Set membership function. Expert and proven need for the use of Gaussian membership functions for variables that have similar values of membership functions of terms. Types defined membership functions for linguistic variables are shown in Table 1.

Table 1 - Types of membership functions defined

The name of the variable	Type of membership function	Range
Location	Triangular	[0 10]
RoomNumber	Triangular	[0 5]
State	Gaussian	[0 10]
Houseroom	Triangular	[0 1]
Distance-to-transport	Triangular	[0 20]
Price	Gaussian	[0 100]

3) Set logical operation on the basis of t-normal: "And Method" – "min"; "Or Method" – "max".

4) Formed base of 35 fuzzy rules whose form is shown in Figure 1.

5) Fuzzy inference made by the algorithm Mamdani: the logical conclusion is organized using a logical minimum (min), composition with operations max, bringing clarity to the centroid method is made.

Graphical view of the output variable depending on the input is shown in Figure 2. This shows the regular price increases with increasing the prestige of the location of the property or to increase the number of rooms.

Figure 1 – Base of fuzzy rules

Estimation of parameters and analysis of the adequacy of model

In the process of analysis should be identified according the received model of the real dependence should be found to improve the model and defined the practical enjoyment of the results achieved.

The adequacy of the constructed model can be determined by examining the remains of the model using appropriate statistical tests. Balances are calculated as the difference between the actual values of the dependent (due) variable y and the values of this variable, calculated using the model:

δ = ((y − ŷ) ÷ y) • 100,

(1)

where ŷ – point forecast value of the object,

y – the true cost estimates of the object.

The experiments yielded the following accuracy: the expectation of error estimates for the data set – 2.51, RMSE – 6.95; mode – 3,2; amount of data in a set (and the number of estimates of real estate) – 114.

Figure 2 – Graphical view of the output variable depending on the input (number of frames: 5, repetitions: 5, duration of frames: 100ms)

Conclusions

A scientific inquiry and analysis in the field of expert systems realtor company. Methods based on the use of methods of mathematical modeling, decision support theory, optimization methods, fuzzy logic, relational database theory, numerical methods and programming. Further actions are determined by the need to develop mathematical and algorithmic models of the functioning of the expert system realtor companies, as well as the development of a software architecture suitable for practical implementation of the system.

Literature

1. Заде Л. Понятие лингвистической переменной и его применение к принятию приближенных решений. – М.: Мир, 1976
2. Леоненков А.В. Нечеткое моделирование в среде MATLAB и fuzzyTECH. – СПб., 2003.
3. Экономика недвижимости: Учебное пособие – Владим. гос. ун-т; Сост.: Д. В. Виноградов, Владимир, 2007. – 136 с. – Электрон. текст. – Режим доступа: http://subschet.ru/subschet.nsf/93eb7fcbec1d!OpenDocument
4. Программа квартиры – учет аренды и управление недвижимостью [Электронный ресурс] / Рентал Макс – Электрон. текст. – Режим доступа к изд.: http://www.rentalmax.ru/rm_manual.pdf
5. Что такое МЛС, или На пути к совершенству [Электронный ресурс] / Э. Светлова // Investor.Kirov.ru Общественно-деловой портал – Электрон. текст. – 17 Январь 2008 – Режим доступа к изд.: http://www.investor.kirov.ru/analit/detail.php?ID=12164
6. Аналитический центр «Индикаторы рынка недвижимости IRN.RU» [Электронный ресурс] / «Индикаторы рынка недвижимости» IRN.RU – Электрон. текст. – Режим доступа к изд.: http://www.irn.ru/acirn/
7. Нечеткая логика в системах управления [Электронный ресурс] / Textreferat Раздел: Логика – 2007-01-21 – Электрон. текст. – Режим доступа: http://www.textreferat.com/referat-1314-1.html
8. Подходы и методы оценки недвижимости [Электронный ресурс] / Уральское Информационное агентство – Электрон. текст. – 19 ноября 2007 – Режим доступа к изд.: http://ocenka-uia.ru/consultations/clauses/6/1/
9. В. Дьяконов, В. Круглов. Математические пакеты расширения MATLAB. Специальный справочник. – Санкт-Петербург: Питер, 2001 – 480 с
10. Нечеткая логика – математические основы [Электронный ресурс] / BaseGroup Labs – Россия, 1995-2008 – Электрон. текст. – Режим доступа: http://www.basegroup.ru/library/analysis/fuzzylogic/math/
11. Нечеткие запросы к реляционным базам данных [Электронный ресурс] / BaseGroup Labs – Россия, 1995-2008 – Электрон. текст. – Режим доступа: http://www.basegroup.ru/library/analysis/fuzzylogic/queries/
12. Нечеткая логика в системах управления [Электронный ресурс] / Textreferat Раздел: Логика – 2007-01-21 10:32:41 – Электрон. текст. – Режим доступа: http://www.textreferat.com/referat-1314-1.html
13. Р. А. Алиев. Управление производством при нечёткой исходной информации, – М.: Энергоатомиздат, 1991. – 240 с.

While writing the given abstract the master's work has not been completed yet. The final date of the work completed is December, 2010. The text of master's work and materials on this topic can be received from the author or her research guide after the indicated date.