Jacinto Mata Vázquez

Discovering Numeric Association Rules via Evolutionary Algorithm

Association rules are one of the most used tools to discover relationships among attributes in a database. Nowadays, there are many efficient techniques to obtain these rules, although most of them require that the values of the attributes be discrete. To solve this problem, these techniques discretize the numeric attributes, but this implies a loss of information. In a general way, these techniques work in two phases: in the first one they try to find the sets of attributes that are, with a determined frequency, within the database (frequent itemsets), and in the second one, they extract the association rules departing from these sets. In this paper we present a technique to find the frequent itemsets in numeric databases without needing to discretize the attributes. We use an evolutionary algorithm to find the intervals of each attribute that conforms a frequent itemset. The evaluation function itself will be the one that decide the amplitude of these intervals. Finally, we evaluate the tool with synthetic and real databases to check the efficiency of our algorithm.

An evolutionary algorithm to discover quantitative association rules from huge databases without the need for an a priori discretization

Association rules are one of the most frequently used tools for finding relationships between different attributes in a database. There are various techniques for obtaining these rules, the most common of which are those which give categorical association rules. However, when we need to relate attributes which are numeric and discrete, we turn to methods which generate quantitative association rules, a far less studied method than the above. In addition, when the database is extremely large, many of these tools cannot be used. In this paper, we present an evolutionary tool for finding association rules in databases (both small and large) comprising quantitative and categorical attributes without the need for an a priori discretization of the domain of the numeric attributes. Finally, we evaluate the tool using both real and synthetic databases.

OBLIC: Classification System Using Evolutionary Algorithm

We present a new classification system based on Evolutionary Algorithm (EA), OBLIC. This tool is an OBLIque Classification system whose function is to induce a set of classification rules no hierarchical from a database or training set. The core of the algorithm is a EA with real-coded and Pittsburgh approach. Each individual is composed by a no fixed classification rules set what split in regions the search space. The fitness of each classification is obtained by means of the exploration of these regions. The result of the tool is the best classification obtained in the evolutionary process.This paper describe and analyze this new method by comparing with other classification systems on UCI Repository databases. We conclude this paper with some observations and future projects.

A Rule-based Method Applied to the Imbalanced Classification of Radiation Toxicity.

This paper describes a rule-based classifier (DEQAR-C), which is set up by the combination of selected rules after a two-phase process. In the first phase, the rules are generated and sorted for each class, and then a selection is performed to obtain a final list of rules. A real imbalanced dataset regarding the toxicity during and after radiation therapy for prostate cancer has been employed in a comparison with other predictive methods (rule-based, artificial neural networks, trees, Bayesian and logistic regression). DEQAR-C produced excellent results in an evaluation regarding several performance measures (accuracy, Matthews correlation coefficient, sensitivity, specificity, precision, recall and F-measure) and by using crossvalidation. Therefore, it was employed to obtain a predictive model using the full data. The resultant model is easily interpretable, combining three rules with two variables, and suggesting conditions that are mostly confirmed by the medical literature.

Annotating Negation in Spanish Clinical Texts

In this paper we present on-going work on annotating negation in Spanish clinical documents. A corpus of anamnesis and radiology reports has been annotated by two domain expert annotators with negation markers and negated events. The Dice coefficient for inter-annotator agreement is higher than 0.94 for negation markers and higher than 0.72 for negated events. The corpus will be publicly released when the annotation process is finished, constituting the first corpus annotated with negation for Spanish clinical reports available for the NLP community.

Deterministic Extraction of Compact Sets of Rules for Subgroup Discovery

This work presents a novel deterministic method to obtain rules for Subgroup Discovery tasks. It makes no previous discretization for the numeric attributes, but their conditions are obtained dynamically. To obtain the final rules, the AUC value of a rule has been used for selecting them. An experimental study supported by appropriate statistical tests was performed, showing good results in comparison with the classic deterministic algorithms CN2-SD and APRIORI-SD. The best results were obtained in the number of induced rules, where a significant reduction was achieved. Also, better coverage and less number of attributes were obtained in the comparison with CN2-SD.

Research and applications: Improving image retrieval effectiveness via query expansion using MeSH hierarchical structure.

Objective We explored two strategies for query expansion utilizing medical subject headings (MeSH) ontology to improve the effectiveness of medical image retrieval systems. In order to achieve greater effectiveness in the expansion, the search text was analyzed to identify which terms were most amenable to being expanded. Design To perform the expansions we utilized the hierarchical structure by which the MeSH descriptors are organized. Two strategies for selecting the terms to be expanded in each query were studied. The first consisted of identifying the medical concepts using the unified medical language system metathesaurus. In the second strategy the text of the query was divided into n-grams, resulting in sequences corresponding to MeSH descriptors. Measurements For the evaluation of the system, we used the collection made available by the ImageCLEF organization in its 2011 medical image retrieval task. The main measure of efficiency employed for evaluating the techniques developed was the mean average precision (MAP). Results Both strategies exceeded the average MAP score in the ImageCLEF 2011 competition (0.1644). The n-gram expansion strategy achieved a MAP of 0.2004, which represents an improvement of 21.89% over the average MAP score in the competition. On the other hand, the medical concepts expansion strategy scored 0.2172 in the MAP, representing a 32.11% improvement. This run won the text-based medical image retrieval task in 2011. Conclusions Query expansion exploiting the hierarchical structure of the MeSH descriptors achieved a significant improvement in image retrieval systems.OBJECTIVE We explored two strategies for query expansion utilizing medical subject headings (MeSH) ontology to improve the effectiveness of medical image retrieval systems. In order to achieve greater effectiveness in the expansion, the search text was analyzed to identify which terms were most amenable to being expanded. DESIGN To perform the expansions we utilized the hierarchical structure by which the MeSH descriptors are organized. Two strategies for selecting the terms to be expanded in each query were studied. The first consisted of identifying the medical concepts using the unified medical language system metathesaurus. In the second strategy the text of the query was divided into n-grams, resulting in sequences corresponding to MeSH descriptors. MEASUREMENTS For the evaluation of the system, we used the collection made available by the ImageCLEF organization in its 2011 medical image retrieval task. The main measure of efficiency employed for evaluating the techniques developed was the mean average precision (MAP). RESULTS Both strategies exceeded the average MAP score in the ImageCLEF 2011 competition (0.1644). The n-gram expansion strategy achieved a MAP of 0.2004, which represents an improvement of 21.89% over the average MAP score in the competition. On the other hand, the medical concepts expansion strategy scored 0.2172 in the MAP, representing a 32.11% improvement. This run won the text-based medical image retrieval task in 2011. CONCLUSIONS Query expansion exploiting the hierarchical structure of the MeSH descriptors achieved a significant improvement in image retrieval systems.