Under the limited number of staff, and small and medium-sized company with little opportunity to learn project management, it is difficult for the staff to carry out projects in the knowledge systemized method. New method and tools for support are needed to prevent project failure in situations where project management knowledge and experience are scarce. Therefore, in this research, we conducted a collection of contents of failed project cases for small and medium-sized company, and analyzed and categorized failure factors. Based on the results, we examined measures to avoid failures as much as possible for small and medium-sized company projects.

　As an evaluation index of academic papers, the citation count of papers has conventionally been used,but in recent years, a method that uses reactions on social media such as altmetrics has been proposed. In this paper, we propose a method to analyze the characteristics of a paper using a general degree that indicates how generally read, in addition to basic indices that indicate the importance of a paper such as number of views on search service and mentions on social media which based on user’s behavior related to papers on the Internet. Actually, using the search log of CiNii Articles for two years and the reference data on social media, we analyze the differences and characteristics of views and mentions to the same paper set and verify the validity of the proposed index.

　In this research, we clarified problems in classifying citations of academic literature by classi fication experiments. After setting up classification categories from the perspectives of significa nce, evaluation, functions, and positions, we assigned those categories to 1,103 citations extrac ted from 30 medical articles. Regarding the perspective of significance, a certain range of sent ences was required for judgment. Regarding the perspective of evaluation, though it was possi ble to judge whether citations (including indirect ones) are positive or negative, the results de pended on the range of sentences to be considered. Regarding the perspective of functions, it was difficult to ensure exclusivity between categories. In addition, it was found that, while in some cases a wider range of sentences improved the accuracy of judgment, in other cases it could not be judged regardless of the range.

　In this study, a direct and an indirect questionnaire survey of Japanese university students were conducted to accurately understand the actual conditions leading to illegal downloading. Two types of questionnaires were prepared, one asking the respondents directly whether they had engaged in any illegal downloading, and the other asking the same question by using an indirect questioning technique, which allowed the respondents to respond to sensitive issues while maintaining their confidentiality. The percentage of respondents indicating that they “had experience” with illegal downloading was 29.1% (using direct questioning) and 57.0% (using indirect questioning), revealing a difference of 27.9% between the two techniques. The results of this study suggest that illegal downloading may be more widespread than is indicated by existing surveys that rely solely on direct questioning.

　To perform mechanical recycling of massive amounts of waste plastics, we developed an identifier oriented on industrial use and based on Raman spectroscopy. For the purpose of designing a practical classification technique that is accurate, fast, and robust against external noise, a simple comparison of spectroscopic peak intensities is considered to be more useful than a complicated multivariate analysis. However, such peak choosing is carried out empirically so far based on expert spectroscopic knowledges. In this report, we successfully demonstrate the classification of three kind plastics using only two peaks that are selected based on support vector machine (SVM) of machine learning; three types of plastic, polypropylene (PP), polystyrene (PS), and acrylonitrile‐butadiene‐styrene copolymer (ABS), are categorized well even in the case of the large artificially generated noise conditions. We could choose automatically a set of two peaks among six sets of four peaks that become candidates in the Raman spectra. Margin values in SVM allow obtaining the decisive information on how to select the peaks in the spectra quantitatively, for distinguishing different types of plastics.

　We proposed a method to identify the edits adding bibliographic references to Wikipedia. The proposed method consists of the following steps. (1) The method extracts the references and matches them to a bibliographic database to build the basic data set. (2) It obtains the full revision history of the page that includes the references from dump data of Wikipedia. It also extracts identifiers and titles for each reference from the basic data set. (3) The method gets the candidate edits adding the references by using the ways, which use either identifiers or titles. (4) The method selects the oldest one as the edit adding the reference. We evaluated the proposed method by using the data set based on DOI links referenced on English Wikipedia. As a result, the accuracy was 93.3% as a whole and over 90% in 20 out of 22 research fields. We showed that the proposed method was able to identify the edits adding bibliographic references at a high accuracy regardless of research fields.

　Various stylometric features have been proposed in the field of authorship identification. In Japanese, for example, morpheme, POS n-grams, and phrase patterns proved to be an effective means of identifying the author. However, it is difficult to extract expression patterns pertaining to functions and usage in sentences, because these stylometric features are tabulated in terms of words, parts-of-speech, and phrases. This study proposes function phrases that represent functions and usages in the text as stylometric features, to prove their effectiveness in author identification. To demonstrate this, we created a corpus of 400 literary works of 20 authors and compared them with the existing stylometric features (morphemes, POS bigrams, particle distribution, and phrase patterns). The results of the study showed that the proposed stylometric features are effective in authorship identification.

　Discovery stories by data have been decreasing in spite of so-called “Data Era”. High throughput approaches taking advantage of such innovative methods as first principles calculation, laboratory automation and machine learning, and data-centric approaches for complexities have been overwhelming “Think slow” modes of enjoying data tasting for discovery. Are we missing important activities? This is a book review of ”Solving the Mysteries of the Solid Elements“ ^［1］ which reminds me how to enjoy tasting of data in a “Think slow” mode for discovery.