Articles by Giancarlo Macchi Jánica

Artificial neural networks are adaptive models that can be used for classification and pattern recognition purposes. ANNs do not differ from standard statistical models. The main difference between ANNs and traditional statistical models is their construction and definition process. In fact ANNs are adaptive in the sense that they can learn. Landscape Archaeology is a research area where the application of ANNs can be very useful. ANNs can be used for Landscape pattern recognition and Settlement systems modeling. This paper illustrate some aspects of the development of new tools and the application of ANNs in a raster GIS environment for archaeological predictive modeling purposes.

The study of the distributive characteristics of settlement patterns is one of the most important topics in the reconstruction of historical landscapes. Unfortunately, advanced knowledge of GIS techniques, as well as quantitative techniques and methods are required in order to achieve positive results in this area. Furthermore, the application of spatial analysis in archaeological research requires a large and expensive investment of software and hardware to accomplish such tasks. In order to address all these issues, this year we started the development of an analytical tool for the study and analysis of spatial distributions. This software, rather than a GIS application, will be conceived as a general utility for data processing. In fact, the main aim of Spatial Analysis Utilities (SAU) is to accomplish in a quick and easy manner the entire quantitative process. The philosophy of the software is to provide conclusive statistical data with just some mouse clicks to non-advanced GIS users.

The purpose of this article is to illustrate the enormous opportunities that the archaeologist is offered in the study of historical landscapes by the process of reconstructing spatial models. This formal process is characterised by the progressive incorporation of social and ecological variables in the form of numerical cartographic surfaces within a given model. In fact, the distortion of social space requires such a quantity, variety and complexity of operations that they could not be realised and synthesised without the help of a computer. Once this process has been completed, it can be exploited in various fields, like the history of roadways and transportation networks, or for a better comprehension of the effects of certain settlement layers on other levels of social occupation of space.