Bruno M. Nogueira

Hierarchical confidence-based active clustering

In this paper, we address the problem of semi-supervised hierarchical clustering by using an active clustering solution with cluster-level constraints. This active learning approach is based on a concept of merge confidence in agglomerative clustering. The proposed method was compared with an un-supervised algorithm (average-link) and a semi-supervised algorithm based on pairwise constraints. The results show that our algorithm tends to be better than the pairwise constrained algorithm and can achieve a significant improvement when compared to the unsupervised algorithm.

Comparing relational and non-relational algorithms for clustering propositional data

Cluster detection methods are widely studied in Propositional Data Mining. In this context, data is individually represented as a feature vector. This data has a natural non-relational structure, but can be represented in a relational form through similarity-based network models. In these models, examples are represented by vertices and an edge connects two examples with high similarity. This relational representation allows employing network-based algorithms in Relational Data Mining. Specifically in clustering tasks, these models allow to use community detection algorithms in networks in order to detect data clusters. In this work, we compared traditional non-relational data-based clustering algorithms with clustering detection algorithms based on relational data using measures for community detection in networks. We carried out an exploratory analysis over 23 numerical datasets and 10 textual datasets. Results show that network models can efficiently represent the data topology, allowing their application in cluster detection with higher precision when compared to non-relational methods.

HCAC: Semi-supervised Hierarchical Clustering Using Confidence-Based Active Learning

Despite their importance, hierarchical clustering has been little explored for semi-supervised algorithms. In this paper, we address the problem of semi-supervised hierarchical clustering by using an active learning solution with cluster-level constraints. This active learning approach is based on a new concept of merge confidence in agglomerative clustering. When there is low confidence in a cluster merge the user is queried and provides a cluster-level constraint. The proposed method is compared with an unsupervised algorithm (average-link) and two state-of-the-art semi-supervised algorithms (pairwise constraints and Constrained Complete-Link). Results show that our algorithm tends to be better than the two semi-supervised algorithms and can achieve a significant improvement when compared to the unsupervised algorithm. Our approach is particularly useful when the number of clusters is high which is the case in many real problems.

Adaptation of FCANN Method to Extract and Represent Comprehensible Knowledge from Neural Networks

Nowadays, Artificial Neural Networks are being widely used in the representation of physical processes. Once trained, the nets are capable to solve unprecedented situations, keeping tolerable errors in their outputs. However, humans cannot assimilate the knowledge kept by these networks, since such knowledge is implicitly represented by their structure and connection weights. Recently, the FCANN method, based in Formal Concept Analysis, has been proposed as a new approach in order to extract, represent and understand the behavior of the process through rules. In this work, it is presented an adaptation of the FCANN method to extract more comprehensible variables relationships, obtaining a reduced and more interesting set of rules related to a predefined domain parameters subset, which provides a better analysis of the knowledge extracted from the neural networks without the necessity of a posteriori implications mining. As case study the approach FCANN will be applied in solar energy system.

Update on chronic nonspecific lower back pain: rehabilitation

1. What is the benefit of acupuncture in the treatment of chronic nonspecific lower back pain? a. The performance of acupuncture combined with other conservative therapies is not more beneficial than applying the same therapies in isolation b. Real acupuncture (deep insertion into acupoints) is superior to sham acupuncture (superficial insertion at sites away from the acupoints) c. Acupuncture is less effective than TENS d. Acupuncture is less effective than massage

Making good choices of non-redundant n-gramwords

A new complete proposal to solve the problem of automatically selecting good and non redundant n-gram words as attributes for textual data is proposed. Generally, the use of n-gram words is required to improve the subjective interpretability of a text mining task, with n ges 2. In these cases, the n-gram words are statistically generated and selected, which always implies in redundancy. The proposed method eliminates only the redundancies. This can be observed by the results of classifiers over the original and the non redundant data sets, because, there is not a decrease in the categorization effectiveness. Additionally, the method is useful for any kind of machine learning process applied to a text mining task.

Recuperação de Dados Ausentes Através de Redes Neurais Artificiais - Estudo de Caso para uma Base de Dados Mercadológica

Resumo— Dados ausentes em bancos de dados são hoje considerados um dos maiores problemas enfrentados na aplicação de Data Mining. No tratamento destes dados é necessário que as características do banco sejam preservadas, ou seja, que não haja informação perdida nem adicionada sem uma análise mais cuidadosa. O objetivo deste trabalho é mostrar como as Redes Neurais Artificiais junto com o conhecimento tácito do especialista no domínio, podem ajudar a recuperar informações dos atributos ausentes. Neste trabalho, esses dois elementos são combinados para recuperar dados ausentes numa base de dados mercadológicos.

Uma análise exploratória de tópicos de pesquisa emergentes em Informática na Educação

The research in Informatics in Education has attracted a growing interest in the last years. For example, in 2012 Brazilian conferences in this research area received their first qualification score in Qualis Capes, indicating the quality and relevance of the national research in Informatics in Education. However, it is still possible to observe a gap between research conducted in Brazil and in other countries. When comparing national and international conferences, we notice a significant difference in terms of the number of published papers and the Qualis score of the conferences. In this sense, in this paper we investigate the current scenario of the research in Informatics in Education. This investigation has been carried out through the usage of Data Mining techniques to obtain topic hierarchies from the published papers in conferences of the area. An exploratory analysis was carried out to identify the main topics in the area. As a result, we provide an overview of the emerging research topics in the educational contexts.

An exploratory analysis of emerging research topics in computer science education

The research in Informatics in Education has attracted a growing interest in the last years. For example, in 2012 Brazilian conferences in this research area received their first qualification score in Qualis Capes, indicating the quality and relevance of the national research in Informatics in Education. However, it is still possible to observe a gap between research conducted in Brazil and in other countries. When comparing national and international conferences, we notice a significant difference in terms of the number of published papers and the Qualis score of the conferences. In this sense, in this paper we investigate the current scenario of the research in Informatics in Education. This investigation has been carried out through the usage of Data Mining techniques to obtain topic hierarchies from the published papers in conferences of the area. An exploratory analysis was carried out to identify the main topics in the area. As a result, we provide an overview of the emerging research topics in the educational contexts.

Poster 375 Clinical Implications of Peripheral Sensitization During Migraine: A Case Report

were treated with fluoroscopically guided lumbar TFESI. Interventions: Patients were administered 80mg DepoMedrol and 1mL of 0.25% bupivicaine at two separate levels and completed the follow-up assessment. Main Outcome Measures: Patient-reported concordant or discordant provocation was assessed during each injection. The primary outcome measure was self-rated percentage of pain reduction from baseline at follow up. Secondary outcome measures were measurements in activity level and daily analgesic consumption. Results: 100% incidence of injection related provocation which was further subclassified as concordant (66%) or discordant (34%). At 2 week post injection follow up, the discordant group achieved a statistically greater decrease in self-reported pain (76%) compared to the concordant group (58%; t 2.1; df (45); P .04). There were no statistically significant differences between concordant and discordant groups with respect to improvements in functional outcome and decreased use of daily oral pain medications. Conclusions: The incidence of provocation was 100%. Both groups achieved significant pain reduction. Concordant provocation did not predict better outcomes. The discordant group had significantly higher self-reported pain reduction in comparison to the concordant group without concomitant functional improvements and reduction in medications. Concordant provocation is not a predictor of response to TFESI.