Mass valuation of properties is important for purposes like property tax, price indices construction, and understanding market dynamics. There are several ways that the mass valuation can be carried out. This paper reviews the conventional MRA and several other advanced methods such as SAR, Kriging, GWR, and MWR. SAR and Kriging are good for modeling spatial dependence while GWR and MWR are good for modeling spatial heterogeneity. The difference between SAR and Kriging is the calculation of weights. Kriging weights are based on the spatial dependence or so called the semi-variogram analysis of the price data whereas the weights in SAR are based on the spatial contiguity between the sample data. MWR and GWR are special types of regression where study region is subdivided into local sections to increase the accuracy of prediction through neutralizing the heterogeneity of autocorrelations. MWR assigns equal weights for observations within a window while GWR uses distance decay functions. The merits and drawbacks of each method are discussed.
Geographical Information Systems (GIS) and three dimensional (3D) World Wide Web (WWW) applications usage are on the rise. The demand for online 3D terrain visualization for GIS data has increased. Current users demand for more complex data which have higher accuracy and realism. This is aided by the emergence of geo-browsers in the market which provide free service and also cater for the commercialized market. Other new technology driving the market is the use of software such as CityGML, Virtual Reality Markup Language (VRML)/ Entensive 3D (X3D), geoVRML, and Keyhole Markup Language (KML). These technologies also play an important role for this new era of online 3D terrain visualization. The aim of this paper is to implement the online 3D terrain visualization for GIS data by using VRML technology and launching the system into three different web servers. The data used for this system are contour data and high resolution satellite image (QUICKBIRD) for Universiti Putra Malaysia (UPM) area. Testing was done only for satellite image overlaid to 3D terrain data. The web servers used in this experiment were the Spatial Research Group Server in UPM, Universiti Utara Malaysia (UUM) web server, and ruzinoor.my web server. The comparison was based on the performance of web servers in terms of accessibility, uploading time, CPU usage, frame rate per second (fps), and number of users. The results from this experiment will be of help and guidance to the developers in finding the right web servers for the best performance on implementing online 3D terrain visualization for GIS data.