Plackett Burman designs are used to estimate main effects because the columns assigned main effects are mutually orthogonal. However, correlations between estimates of a main effect column and a two-factor interaction column are not orthogonal in general. There are variations in the magnitude of correlation depending on the size of Plackett Burman design. For example, correlation coefficients between a main effect and a two-factor interaction are either 0 or ±0.33 in Plackett Burman design N=12. However, most of them are ±0.2 and some of them are 0.6 in Plackett Burman design N=20. When a correlation coefficient between main effects and two-factor interactions is significantly large, analysis results may be depart from the true status. Therefore, in this study, we comprehensively describe correlation coefficients between a main effect and a two-factor interaction for all designs proposed by the Plackett and Burman designs such that N=8,...,100. In addition, for designs that have large correlation coefficients, they are classified according to how to construct a design table and examined in detail. We describe certain rules in combination of a main effect and a two-factor interaction with large correlation coefficient. We also show how to construct designs that avoids large correlation based on the rules.
As the information of vehicle defect, the vehicle defect information and the recall information are disclosed by the website of Japanese Ministry of Land, Infrastructure, Transport and Tourism., Unlike other studies which analyze the text data of each one, we focus on the relation between them and analyze it of both. In order to derive the tendency of dependency pairs which means defect phenomena, dependency parsing and BM 25 are applied for the analysis. From results of calculating occurrence and BM 25 score for pairs of dependency words and modified words, some tendencies are found in each of high score's pairs. For the analysis of recall information, correspondence analysis, co-occurrence network and self-organizing map are conducted to find the relational words between three categories of defect factor. Twelve primary factors of five parts are shown as the factors in fuel device which causes liquid leakage. Finally, factors of important defect phenomena is discussed. Several dependency pairs are shown as important defect phenomena, and primary factors of parts are shown as an example of factors of liquid leakage which is one of them.
Generally, statistical learning using sufficient training data enables a highly accurate classification. However, it is sometimes difficult to collect sufficient training data for constructing an accurate classifier. In particular, for classification problems, correct labels corresponding to each feature vector are required. Therefore, semi supervised learning that uses not only labeled training data but also a large amount of unlabeled data for acquiring an accurate classifier has recently received attention. In a semi-supervised learning setting, if the distribution of labeled data is biased in each category set, it is difficult to estimate the correct labeling for unlabeled data. Consequently, the classification accuracy is degraded owing to the incorrect labeling. SemiBoost is a type of semi supervised learning method that avoids the above problem and has high performance. However, this method is a binary classification method and cannot be extended directly to handle multi-valued classification problems. In this paper, we propose a method to extend SemiBoost to enable it to perform multi-valued classification by introducing to it the concept of the error correcting output code (ECOC) method. Using the proposed method enables a more accurate labeling for unlabeled data. To verify the effectiveness of our proposed method, we conducted simulation experiments by using the data from the UCI Machine Learning Repository. The experimental results showed that the proposed method is effective for biased data. In addition, the classification results when the ratio of bias data was changed are shown and discussed.