Koichiro Iwata

Discovery of Sequential Patterns Coinciding with Analysts’ Interests

This paper proposes a new sequential pattern mining method. The method introduces a new evaluation criterion satisfying the Apriori property. The criterion is calculated by the frequency of the sequential pattern and the minimum frequency of items included in the items. It extracts sequential patterns that can be rules predicting future items with high probability. Also, the method introduces new constraints. The constraints extract item sets composed of items whose attributes are different and extracts sequential patterns composed of item sets whose attribute sets are equal to one another. The proposed method efficiently discovers sequential patterns coinciding with analysts’ interests by combining the criterion and the constraints. The paper verifies the effectiveness of the proposed method by applying it to medical examination data.

Prioritizing Health Promotion Plans with k-Bayesian Network Classifier

Recently, Bayesian network classifiers (BNCs) have attracted many researchers because they can produce classification models with dependencies among attributes. From the application viewpoint, however, BNCs sometimes produce models too complicated to interpret easily. In this paper, we propose k-Bayesian network classifier (k-BNC), which is a new method to reconstruct the attribute-dependency relationship from data for health promotion planning. From the health promotion viewpoint, it would be highly advantageous if occupational physicians could make effective plans for employees, and if employees could carry out the plans easily. Therefore, we focus on the attribute dependencies in classification models represented as a directed acyclic graph (DAG), and find the effective attributes by measuring the standardized Kullback-Leibler divergence from parent attributes to their children. In experimental evaluation, we firstly compare the accuracy of k-BNC with that of Naive Bayes Classifiers, and other wellknown Bayesian Networks and structure learning methods (k2 algorithm etc.) on some public datasets. We show that our proposed k-BNC method successfully produces classification models for the prioritization of health promotion plans on our health checkup data.

Genetic Algorithm Based Methods for Identification of Health Risk Factors Aimed at Preventing Metabolic Syndrome

In recent years, metabolic syndrome has emerged as a major health concern because it increases the risk of developing lifestyle diseases, such as diabetes, hypertension, and cardiovascular disease. Some of the symptoms of the metabolic syndrome are high blood pressure, decreased HDL cholesterol, and elevated triglycerides (TG). To prevent the developing of metabolic syndrome, accurate prediction of the future values of these health risk factors and identification of other factors from the health checkup and lifestyle data, which are highly related with these risk factors, are very important. In this paper, we propose a new framework, based on genetic algorithm and its variants, for identifying those important health factors and predicting the future health risk of a person with high accuracy. We show the effectiveness of the proposed system by applying it to the health checkup and lifestyle data of Toshiba Corporation.

Instability of the NS5A ISDR of hepatitis C virus during natural course : take-over of wild type by mutant type or vice-versa driven by immune pressure

Abstract The so-called interferon sensitivity determining region (ISDR) in the NS5A protein of hepatitis C virus (HCV) undergoes mutational changes in chronically infected hosts, but its mechanism remains obscure. We analyzed the ISDR by direct sequencing and/or restriction fragment length polymorphism in serial sera from two patients with chronic hepatitis C. Patient 1 showed a change in the ISDR quasispecies from wild type to mutant type, while in patient 2 the situation was vice-versa, during a period where interferon was not administered. The emerging mutant in patient 1 was detected in IgG-unbound HCV fraction, while the fading-out mutant in patient 2 was IgG-bound. These results suggest that an immune pressure against virion surface epitope(s) plays an insidious role for the apparent selection of HCV ISDR quasispecies. Of note, the ISDR mutants found in the two patients, irrespective of IgG-bound or -unbound, disappeared promptly by interferon therapy.

Risk prediction and risk factors identification from imbalanced data with RPMBGA

In this paper, we propose a new method to predict the risk of an event very accurately from imbalanced data in which the number of instances of the majority class is very larger than that of the minority class and to identify the features that are relevant for the target risk factor. To solve the trade-off between the prediction rates of the majority and the minority classes, three input parameters are used, which supply the costs of misclassification of an instance from the majority and the minority classes or the sensitivity threshold of the minority class. To get relevant features and to utilize the prior information about the relationship of a feature with the target risk factor, a probabilistic model building genetic algorithm called RPMBGA+ is employed. By applying the proposed technique to the health checkup and lifestyle data of Toshiba Corporation, we have found that the proposed method improves the sensitivity of the minority class and selects a very small number of informative features.