• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Avalia??o em massa de im?veis rurais atrav?s de modelagem cl?ssica, espacial e geoestat?stica / Mass appraisal of rural land through classical, spatial and geostatistics modeling

UBERTI, Marlene Salete 11 July 2016 (has links)
Submitted by Jorge Silva (jorgelmsilva@ufrrj.br) on 2018-05-18T17:57:14Z No. of bitstreams: 1 2016 - Marlene Salete Uberti.pdf: 4547445 bytes, checksum: 0815ad8e3c8b5cdf64206ec60d91619c (MD5) / Made available in DSpace on 2018-05-18T17:57:14Z (GMT). No. of bitstreams: 1 2016 - Marlene Salete Uberti.pdf: 4547445 bytes, checksum: 0815ad8e3c8b5cdf64206ec60d91619c (MD5) Previous issue date: 2016-07-11 / Traditionally the Classical Linear Regression Models (CLRMs) have been used for mass appraisal of property bulk value, however, it has been noticed the need to take into account the data spatial variation. This modeling for the geographic effects has been used mainly in urban area appraisals, while farmland values are also affected by geographic location. The lack of methodologies for mass evaluation of farmland has led to tax evasion of farmland tax revenue (ITR), as it has been inefficiently and inexpressively collected since its enactment in 1964. The objective of this paper is to use econometrics models of spatial regression in farmland comparables to produce a map of standard ground value for the Northern Region of Rio de Janeiro State, Brazil. The proposed methodology includes the investigation and modeling the effects of spatial autocorrelation on the CLRMs, to evaluate their performance comparing them with the spatial models and to produce a map of standard ground value through ordinary Kriging and kernel estimator. The sample of comparables was comprised of 113 observations for model development and 25 observations for validation. To evaluate the performance of obtained maps of values were used the validation samples to calculate the Root Mean Square Error (RMSE) values and the metrics recommended by the International Association of Assessing Officers (IAAO). The results showed that the spatial autocorrelation can have its effect predicted by the Conditional AutoRegressive model (CAR) and by the Geographically Weighted Regression (GWR). By using the values predicted with the GWR model and the validation comparables, the Kernel estimator presented the best performance on map production, yielding the lowest RMSE and dispersion coefficients, median of ratios and Price Related Differential (PRD) close to IAAO recommended values. The combination of classical and spatial regression methodologies and the use of Geostatistics techniques showed to be suitable for obtaining maps of standard ground value for farmland areas. The proposed methodology has been show applicable to farmland sales market, as it can be used by municipalities to obtain representative models of real market values, as well as to produce farmland standard ground value maps. / Nas avalia??es em massa de im?veis tradicionalmente s?o utilizados os modelos cl?ssicos de regress?o linear (MCRL), entretanto tem-se verificado a necessidade de modelar os dados espacialmente. Esta modelagem dos efeitos espaciais vem sendo utilizada principalmente nas avalia??es de ?reas urbanas, sendo que os valores dos im?veis nas ?reas rurais tamb?m s?o afetados pela localiza??o geogr?fica. A inexist?ncia de metodologias de avalia??o em massa de im?veis rurais ? um dos motivos da evas?o da receita do imposto territorial rural (ITR), pois desde que foi criado em 1964, a arrecada??o deste imposto ? ineficiente e inexpressiva. O objetivo deste trabalho foi a utiliza??o de modelos econom?tricos de regress?o espacial na modelagem dos efeitos espaciais em uma amostra de im?veis rurais para a elabora??o da Planta de Valores Gen?ricos (PVG) em uma ?rea da Regi?o Norte Fluminense, estado do Rio de Janeiro. A proposta metodol?gica consistiu em investigar e modelar os efeitos causados pela autocorrela??o espacial sobre os MCRL, avaliar seus desempenhos comparando-os com os modelos espaciais e produzir a PVG por meio da Krigagem ordin?ria e do estimador Kernel. A amostra utilizada contou com 113 observa??es e 25 amostras de verifica??o. Para avaliar o desempenho das superf?cies de valores obtidas foram utilizadas as amostras de verifica??o e calculados os valores da Raiz Quadrada do Erro M?dio Quadr?tico (REMQ) e das m?tricas recomendadas pela International Association of Assessing Officers (IAAO). Os resultados mostraram que a autocorrela??o espacial pode ter seus efeitos reduzidos pelo Modelo do Erro Espacialmente Correlacionado (Conditional Auto Regressive - CAR) e pela Regress?o Geograficamente Ponderada (RGP). A superf?cie gerada pelo estimador Kernel, utilizando-se os valores preditos da amostra de verifica??o pelo modelo RGP foi a que obteve o melhor desempenho com menor REMQ e valores do coeficiente de dispers?o (COD), da mediana das raz?es e do Diferencial Relativo ao Pre?o (Price Related Differential - PRD) pr?ximos dos recomendados pela IAAO. A combina??o das metodologias da regress?o cl?ssica e espacial, e a utiliza??o de t?cnicas de Geoestat?stica se mostraram adequadas para a elabora??o e obten??o da PVG para ?reas rurais. A metodologia proposta se mostrou aplic?vel nos mercados de terras rurais, pois pode ser utilizada pelos munic?pios para obter modelos representativos da realidade destes mercados, bem como para elaborar a PVG das ?reas rurais.

Page generated in 0.0762 seconds