 |  || DonNTU > Master's Portal |
Computer Sciences and Technologies Faculty |
|
|
|
|
Sergey Nechaev |
|
Faculty: Pulpit: Speciality: Theme of master's work: Scientific Supervisor: |
Computer Sciences and Technologies Faculty Automated Control Systems Information Control Systems and Technologies аssociate professor of Automated Control Systems department Tamara P. Zhukova |
|
Abstract
of the qualification master’s work
«DEVELOPMENT OF THE COMPUTER SYSTEM EVALUATION OF PROPERTIES REAL ESTATE OBJECTS WITH USING OF NEURAL NETWORKS»
Supervisor of the Scientific Work: candidate of phys-math. sciences, an associate professor Tamara P. Zhukova,
Donetsk National Technical University,
associate professor of Automated Control Systems department
Donetsk National Technical University,
associate professor of Automated Control Systems department
- RELEVANCE
- PURPOSE AND TASKS OF DESIGH AND RESEARCH
- SCIENTIFIC NOVELTY
- PRACTICAL SIGNIFICANCE OF RESULTS
- SUMMARY
REFERENCES
INTRODUCTION
Property valuation is one of the most popular and at the same time, one of the most difficult tasks in the market of intelligent systems for the assessment and decision making. The difficulty is, firstly, a large number of factors influencing the evaluation. Secondly, the nature of the factors represents a significant challenge — some of them quite difficult to formalize (eg, "the degree of prestige area of finding the object," the appearance of the object, information about the history of the subject being assessed, the analysis of the location of the object, etc.). Third, the real estate market is fairly dynamic, implying a high rate of change of the parameters estimates over time. Fourthly, to create training samples and knowledge bases have to use the experience of different assessors, which may lead to conflicting decisions.
RELEVANCE
The relevance of this system justified by the fact that at that time real estate is one of the main niches in the economy. Because of this convenience, speed and ease of search with the help of a computerized system that will help a person objectively assess the value of a property without a realtor.
And as neural networks based on the primitive biological model of nervous systems and are capable of learning, the assessment of real estate by the National Assembly looks very attractive.
Another relevant factor is the independence from realtor. At the moment, property is assessed — man (realtor). Realtor may not always be objective in their assessment, and various properties with approximately the same performance can be assessed in different results. It is therefore advisable to use a system that can operate without interference.
PURPOSE AND TASKS OF DESIGH AND RESEARCH
The purpose of this study is the creation of an automated system of property valuation, with the help of neural networks. In accordance with the purpose, identified the following objectives:
Currently, most real estate is estimated by hand — a man. To improve efficiency, speed and reliability of assessment, it is advisable to use computer power to carry out this work. That is, the novelty lies in the fact that the assessment of the property, will take place without human intervention. Do not be deliberately inflated prices, the possibility to implement this system on the Web portal or, for information terminals of the city.
PRACTICAL SIGNIFICANCE OF RESULTS
Experiment has been conducted and it was found that using this approach, we can obtain good results, providing reliable estimates of 73-85% of cases. This approach provides better results than the fuzzy evaluation system property FuzzyExtent. Significant deviations in the evaluation revealed only 2% of cases. A system as a whole, will help users to get quick and not depend on an assessment of the realtor or other real estate.
SUMMARY
In the course of research work have been analyzed in this study, scientists from Russia and the UK. Was expanded body of knowledge about real estate and neural networks. A representation of the factors that may be applied to evaluate the performance of real estate. A general structure of the system of property valuation, with the use of neural networks.
REFERENCES
1. Журнал "Приборы и системы. Управление, контроль, диагностика"/Применение нейронных сетей для оценки характеристик недвижимости, Пителинский К.В., Тюркин А.А., 2008г
2. Нейронные сети: Учебное пособие, Беркинблит М.Б. — М.: МИРОС и ВЗМШ РАО, 1993. — 96 с: ил.
3. Обучение нейронных сетей, Горбань А.Н., : СП «Параграф», 1990.- 154с.
4. Особенности нейросетевого моделирования в задаче массовой оценки муниципальной недвижимости г. Москвы, К.К. Борусяк, И.В. Мунерман «Институт управления стоимостью»
5. Основные концепции нейронных сетей, Каллан Роберт: Пер. с англ. – М.: Издательский дом «Вильямс», 2001 – с.:ил – Парал. тит. англ.
6. Организация и использование нейронных сетей (методы и технологии) Аксенов С.В., Новосельцев В.Б./Под общ. ред.В.Б. Новосельцева. – Томск: Изд-во НТЛ, 2006. – 128 с.
7. Информационно образовательный ресурс кафедры "Информационные технологии"/Финансовая академия при Правительстве Российской федерации.
[Электронный ресурс] — Режим доступа: http://fakit.narod.ru/ai.mht
8. Электронный учебник по статистике. StatSoft, Inc. (2001). Москва, StatSoft.
[Электронный ресурс] — Режим доступа: http://www.statsoft.ru/home/textbook/default.htm раздел нейронные сети
9. Информационно познавательный журнал «Виктория», тема «Нейросети в задачах отображения»
[Электронный ресурс] — Режим доступа: http://www.victoria.lviv.ua/html/oio/html/theme8_rus.htm
While writing the given abstract the master's work has not been completed yet. The final date of the work completed is December, 1st, 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.
Property valuation is one of the most popular and at the same time, one of the most difficult tasks in the market of intelligent systems for the assessment and decision making. The difficulty is, firstly, a large number of factors influencing the evaluation. Secondly, the nature of the factors represents a significant challenge — some of them quite difficult to formalize (eg, "the degree of prestige area of finding the object," the appearance of the object, information about the history of the subject being assessed, the analysis of the location of the object, etc.). Third, the real estate market is fairly dynamic, implying a high rate of change of the parameters estimates over time. Fourthly, to create training samples and knowledge bases have to use the experience of different assessors, which may lead to conflicting decisions.
RELEVANCE
The relevance of this system justified by the fact that at that time real estate is one of the main niches in the economy. Because of this convenience, speed and ease of search with the help of a computerized system that will help a person objectively assess the value of a property without a realtor.
And as neural networks based on the primitive biological model of nervous systems and are capable of learning, the assessment of real estate by the National Assembly looks very attractive.
Another relevant factor is the independence from realtor. At the moment, property is assessed — man (realtor). Realtor may not always be objective in their assessment, and various properties with approximately the same performance can be assessed in different results. It is therefore advisable to use a system that can operate without interference.
PURPOSE AND TASKS OF DESIGH AND RESEARCH
The purpose of this study is the creation of an automated system of property valuation, with the help of neural networks. In accordance with the purpose, identified the following objectives:
- Determining the number of factors affecting the real estate appraisal
- Collecting all data (factors) real estate
- Formalization of some of the factors for their application
- Providing high-update, due to the dynamic real estate market
- Defining the parameters of the neural network
- Creating a software product
Currently, most real estate is estimated by hand — a man. To improve efficiency, speed and reliability of assessment, it is advisable to use computer power to carry out this work. That is, the novelty lies in the fact that the assessment of the property, will take place without human intervention. Do not be deliberately inflated prices, the possibility to implement this system on the Web portal or, for information terminals of the city.
PRACTICAL SIGNIFICANCE OF RESULTS
Experiment has been conducted and it was found that using this approach, we can obtain good results, providing reliable estimates of 73-85% of cases. This approach provides better results than the fuzzy evaluation system property FuzzyExtent. Significant deviations in the evaluation revealed only 2% of cases. A system as a whole, will help users to get quick and not depend on an assessment of the realtor or other real estate.
SUMMARY
In the course of research work have been analyzed in this study, scientists from Russia and the UK. Was expanded body of knowledge about real estate and neural networks. A representation of the factors that may be applied to evaluate the performance of real estate. A general structure of the system of property valuation, with the use of neural networks.
REFERENCES
1. Журнал "Приборы и системы. Управление, контроль, диагностика"/Применение нейронных сетей для оценки характеристик недвижимости, Пителинский К.В., Тюркин А.А., 2008г
2. Нейронные сети: Учебное пособие, Беркинблит М.Б. — М.: МИРОС и ВЗМШ РАО, 1993. — 96 с: ил.
3. Обучение нейронных сетей, Горбань А.Н., : СП «Параграф», 1990.- 154с.
4. Особенности нейросетевого моделирования в задаче массовой оценки муниципальной недвижимости г. Москвы, К.К. Борусяк, И.В. Мунерман «Институт управления стоимостью»
5. Основные концепции нейронных сетей, Каллан Роберт: Пер. с англ. – М.: Издательский дом «Вильямс», 2001 – с.:ил – Парал. тит. англ.
6. Организация и использование нейронных сетей (методы и технологии) Аксенов С.В., Новосельцев В.Б./Под общ. ред.В.Б. Новосельцева. – Томск: Изд-во НТЛ, 2006. – 128 с.
7. Информационно образовательный ресурс кафедры "Информационные технологии"/Финансовая академия при Правительстве Российской федерации.
[Электронный ресурс] — Режим доступа: http://fakit.narod.ru/ai.mht
8. Электронный учебник по статистике. StatSoft, Inc. (2001). Москва, StatSoft.
[Электронный ресурс] — Режим доступа: http://www.statsoft.ru/home/textbook/default.htm раздел нейронные сети
9. Информационно познавательный журнал «Виктория», тема «Нейросети в задачах отображения»
[Электронный ресурс] — Режим доступа: http://www.victoria.lviv.ua/html/oio/html/theme8_rus.htm
While writing the given abstract the master's work has not been completed yet. The final date of the work completed is December, 1st, 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.
Materials on the topic of final work: Autobiography | Abstract