Alberto Gemelli

Feature Ranking Based on Decision Border

In this paper a Feature Ranking algorithm for classification is proposed, which is based on the notion of Bayes decision border. The method elaborates upon the results of the Decision Border Feature Extraction approach, exploiting properties of eigenvalues and eigenvectors of the orthogonal transformation to calculate the discriminative importance weights of the original features. Non parametric classification is also considered by resorting to Labeled Vector Quantizers neural networks trained by the BVQ algorithm. The choice of this architecture leads to a cheap implementation of the ranking algorithm we call BVQ-FR. The effectiveness of BVQ-FR is tested on real datasets. The novelty of the method is to use a feature extraction technique to assess the weight of the original features, as opposed to heuristics methods commonly used.

A Feature Ranking Component for GIS Architecture

For a Geographical Information System, it is wanted to design a permanent architectural component which is committed to select relevant information from large databases and network of sensors, and useful in the wide range of applications that such system is destined to. This component allows the system to reach an optimal performance with savings on data acquisition costs and computational resources. The component is based on Data Mining technology and uses feature extraction algorithms to rank the relevance of the dataset features. In all applications reducible to a classification of geographical objects, the feature ranking procedure highlights the features with higher class discriminatory power. Features Ranking is also a cognitive strategy, and produces models of interest toward an artificial intelligent system. The GIS is a Decision Support System (DSS), whereas the feature ranking component supports a within processing decisional activity. A prototype of this component has been assembled and tested on several geographical dataset with promising results at current research stage.

GIS-Supported Decision Making

For the regional scale planning of the renewable energy production, it is required to systematically assess the availability of the resource and the convenience of its exploitation in each site individually. The GIS is the ideal tool to process large amounts of data that subtend the decision problem in a geographic dimension, to synthesize the models, and make them accessible. In this chapter, we present some analytical tools that are included in the architecture of GIS and extend its capabilities as a decision support system. The emphasis is on methods that automate issues such as selection of information, site classification, problems solving. An information system is outlined that has the ability to self-configure to maximize efficacy and efficiency of the data processing.

A goal oriented feature selection for collaborative GIS

Geographical Information Systems, in the modern conception, are open and collaborative systems. In this framework, we present a novel goal oriented feature selection method. The method ranks and selects the information on the basis of specific classification and costs optimization targets, provided by the user. The output models is human comprehensible and can be used jointly with others rank models, supporting a multi-goals analysis. A case study is also presented.

A Geometric Approach to Feature Ranking Based Upon Results of Effective Decision Boundary Feature Matrix

This chapter presents a new method of Feature Ranking (FR) that calculates the relative weight of features in their original domain with an algorithmic procedure. The method supports information selection of real world features and is useful when the number of features has costs implications. The Feature Extraction (FE) techniques, although accurate, provide the weights of artificial features whereas it is important to weight the real features to have readable models. The accuracy of the ranking is also an important aspect; the heuristics methods, another major family of ranking methods based on generate-and-test procedures, are by definition approximate although they produce readable models. The ranking method proposed here combines the advantages of older methods, it has at its core a feature extraction technique based on Effective Decision Boundary Feature Matrix (EDBFM), which is extended to calculate the total weight of the real features through a procedure geometrically justified. The modular design of the new method allows to include any FE technique referable to the EDBFM model; a thorough benchmarking of the various solutions has been conducted.

GIS-Supported Decision Making for Low-Temperature Geothermal Energy in Central Italy

The project specifications of a geospatial decision support system are described in this chapter. The system is dedicated to the low-temperature geothermal energy, and it allows to build spatial models of the resource and to evaluate the cost-benefit of its exploitation in a wide region of central Italy. The architecture of the system and the computational workflow are depicted. The design is based on a thorough study of data analysis tools.

Decision Environment of Renewable Energy: The Case of Geothermal Energy

The exploitation of low temperature geothermal energy (LTGE) for thermoregulation is an expanding activity with various applications. The production of LTGE, although it is an efficient process, it has a cost benefit varying from one site to another depending, but not only, of natural factors. For local administrations and for those who invest in the provision of energy services, a regional model of the distribution of the LTGE resource is necessary for planning production, incentives, and investment. Also, the factors that influence the cost benefit of the resource must be studied in the spatial dimension. The construction of spatial models is a process requiring the acquisition of large amounts of data, the use of computer technology, and a substantial process design effort. In this chapter, the emphasis is placed on the support of Geographic Information System (GIS) in spatial modeling of LTGE cost benefit. The techniques for collecting spatial data on a large geographic scale are introduced. The technical and organizational aspects discussed delineate an information environment aimed to providing decision support in the regional development of LTGE.

Feature Analysis: Selecting Decision Criteria

The data acquired constitute a wealth of information still to be reduced and transformed for the specific goals of the decision problem. In this chapter, the automated reasoning techniques are used to maximize the efficiency of data processing, and to extract indicators of the LTGE potential and of the energy demand, this resource must meet. The final products of this process are indicators of performance of the LTGE plant and other alternative systems. The cost-benefit comparison of the plants is referred to a site chosen as reference and to standardized production.

Regional Atlas Supporting the Decision-Making Process

In this chapter, a set of cartographic products has been gathered to support the decisions in the case study. Three different methods of analysis are highlighted, based on maps that represent different levels of support: (1) passive level, maps of individual features; (2) intermediate level, maps in which the decision criteria are represented; (3) active level, maps that express directly the solution to the decision problem. The collection constitutes a goal-oriented decision atlas dedicated to LTGE.

Decision Analysis: Choosing the Right Plant

A series of decision problems concerning the choice of the optimal plant are defined and solved in each cell of the geographic grid. To this end, the cost-benefit analysis of the geothermal resource has been extended to the geographic dimension. The decision analysis is conducted on the basis of the decision criteria and priorities for different categories of stakeholders. The goal is to automate the decision problem solving, to build spatial decision-making models and represent them in map format.