Browsing by Author "Merlino, Hernán Daniel"

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Co-location rules discovery process focused on reference spatial features using decision tree learning
    (2017) Rottoli, Giovanni Daián; Merlino, Hernán Daniel; García Martínez, Ramón
    The co-location discovery process serves to find subsets of spatial features frequently located together. Many algorithms and methods have been designed in recent years; however, finding this kind of patterns around specific spatial features is a task in which the existing solutions provide incorrect results. Throughout this paper we propose a knowledge discovery process to find co-location patterns focused on reference features using decision tree learning algorithms on transactional data generated using maximal cliques. A validation test of this process is provided.
  • Thumbnail Image
    Item
    Knowledge discovery process for description of spatially referenced clusters
    (2017) Rottoli, Giovanni Daián; Merlino, Hernán Daniel; García Martínez, Ramón
    Spatial clustering is an important field of spatial data mining and knowledge discovery that serves to partition a spatial data set to obtain disjoint subsets with spatial elements that are similar to each other. Existing algorithms can be used to perform three types of cluster analyses, including clustering of spatial points, regionalization and point pattern analysis. However, all these existing methods do not provide a description of the discovered spatial clusters, which is useful for decision making in many different fields. This work proposes a knowledge discovery process for the description of spatially referenced clusters that uses decision tree learning algorithms. Two proofs of concept of the proposed process using different spat ial clustering algorithm on real data are also provided.
  • Thumbnail Image
    Item
    Knowledge discovery process for detection of spatial outliers
    (2018) Rottoli, Giovanni Daián; Merlino, Hernán Daniel; García Martínez, Ramón
    Detection of spatial outliers is a spatial data mining task aimed at discovering data observations that differ from other data observations within its spatial neighborhood. Some considerations that depend on the problem domain and data characteristics have to be taken into account for the selection of the data mining algorithms to be used in each data mining project. This massive amount of possible algorithm combinations makes it necessary to design a knowledge discovery process for detection of local spatial outliers in order to perform this activity in a standardized way. This work provides a proposal for this knowledge discovery process based on the Knowledge Discovery in Database process (KDD) and a proof of concept of this design using real world data.