José Francisco Martínez Trinidad

Learn More
The so-called logical combinatorial approach to Pattern Recognition is presented, and works (mainly in Spanish and Russian) that are not ordinarily available, are exposed to the Western reader. The use of this approach for supervised and unsupervised pattern recognition, and for feature selection is reviewed. Also, an uni"ed notation describing the original(More)
In supervised learning, a training set providing previously known information is used to classify new instances. Commonly, several instances are stored in the training set but some of them are not useful for classifying therefore it is possible to get acceptable classification rates ignoring non useful cases; this process is known as instance selection.(More)
In supervised classification, a training set T is given to a classifier for classifying new prototypes. In practice, not all information in T is useful for classifiers, therefore, it is convenient to discard irrelevant prototypes from T. This process is known as prototype selection, which is an important task for classifiers since through this process the(More)
reuse any copyrighted component of this work in other works must be obtained from the IEEE. " Abstract Burrows-Wheeler transform (BWT) has received special attention due to its effectiveness in lossless data compression algorithms. However, implementations of BWT-based algorithms have been limited due to the complexity of the suffix sorting process applied(More)
In this paper, two algorithms for discovering all the Maximal Sequential Patterns (MSP) in a document collection and in a single document are presented. The proposed algorithms follow the " pattern-growth strategy " where small frequent sequences are found first with the goal of growing them to obtain MSP. Our algorithms process the documents in an(More)
In this work a new set of clustering criteria for the structuralization of universes in the logical combinatory approach of the Pattern Recognition Theory is introduced. The criteria are considered for classical partitions and covers as well as fuzzy classes. The existing relationships between these criteria are studied. An example of an application of the(More)
Support calculation and duplicate detection are the most challenging and unavoidable subtasks in frequent connected subgraph (FCS) mining. The most successful FCS mining algorithms have focused on optimizing these subtasks since the existing solutions for both subtasks have high computational complexity. In this paper, we propose two novel properties that(More)