著者
財津 亘 渋谷 友祐 長谷川 直宏
出版者
日本法科学技術学会
雑誌
日本法科学技術学会誌 (ISSN:18801323)
巻号頁・発行日
vol.13, no.1, pp.83-92, 2008 (Released:2008-04-19)
参考文献数
28

Offender profiling is one of the tools of decision making for criminal investigation. It is a set of techniques to infer characteristics of an unknown offender, such as sex, age bracket, lifestyle, psychological feature, previous crime, inhabited area, from the information which is left at the crime scene.   In this article, we proposed a tool of decision-making for criminal investigation from the perspective of prediction of an uncertain event by the use of a Bayesian Network (BN). BN is a probability model that describes causal structure of events as chain networks of conditional probability, and is capable to predict the possibility of uncertain events.   To examine the validity of the constructed model, firstly, we divided previous offenders’ information of the indoor-sex-offence cases into a training data (9,859 cases) and validation data (50 cases). Secondly, we constructed a model from the training data by means of K2 and MDL (minimum description length) as search-algorithm and information criteria, respectively. Finally, the validity of the model was examined by the validation data as virtual cases.   According to the model, 21 target variables (16 behavioral variables, 2 vehicle variables and 3 victim variables) linked the explanatory variable (employment) directly, and most of these variables related to the employment. The results of the model validity showed that the accuracy of predicting the employment increased 10% higher when the age bracket could be estimated from the testimony of the victim.   The results indicated that the BN model of the offender profiling would be able to provide valuable information for decision making for crime investigation. To predict characteristics of an unknown offender more accurately, it is crucial to select more appropriate information criteria and develop the search-algorithm, as well as to construct the database from more accurate information.
著者
財津 亘
出版者
日本犯罪心理学会
雑誌
犯罪心理学研究 (ISSN:00177547)
巻号頁・発行日
vol.56, no.1, pp.77-88, 2018-08-27 (Released:2018-09-19)
参考文献数
21

本研究では,性犯罪者519名が敢行した性犯罪の罪種間における移行性やそれら性犯罪と各種窃盗犯罪との関連について検討することを目的とした。性犯罪者519名が敢行して検挙された各種性犯罪ならびに各種窃盗の犯罪経歴を基に,データセット(519名×15犯罪種別)を作成し,分析を行った。対応分析ならびに階層的クラスター分析の結果によると,「犯行場所」と「身体的接触の有無」の2次元が抽出され,性犯罪は,それら2つの次元に沿って①接触型(屋内強姦,屋外強姦,強制わいせつ,年少者わいせつ),②非接触型(露出,のぞき),③窃盗型(色情盗)に分類された。また,窃盗犯罪は,①侵入窃盗(空き巣,忍込み,出店荒し),②乗物盗(自動車盗,オートバイ盗,自転車盗),③非侵入窃盗(車上ねらい,万引き)に分類された。さらには,強制わいせつから屋外強姦,露出から年少者わいせつへの移行性が示唆された。そのほかにも,接触型性犯罪と乗物盗の間に相対的に関連性がみられたほか,屋内強姦は,侵入窃盗との関連がみられた。非接触型性犯罪は,窃盗犯罪との関連が低いことを示唆した。
著者
財津 亘
出版者
日本犯罪心理学会
雑誌
犯罪心理学研究 (ISSN:00177547)
巻号頁・発行日
vol.53, no.2, pp.29-41, 2016-01-31 (Released:2017-03-23)
参考文献数
22

本研究は,最近10年間の放火事件を対象に,単一ならびに連続放火犯が述べた動機をテキストマイニングによって分類することを目的とした。研究1では,まず単一放火犯が供述した動機に関する文章(文字列)を対象に名詞を抽出し,放火犯ごとで名詞の出現度数を算出した。これにより作成されたデータセット(253名×67名詞)について,古典的多次元尺度法(重み付きユークリッド距離)による分析を行い,名詞を2次元上に布置した。つづいて,2次元上に布置された名詞の座標データを基に,階層的クラスター分析(Ward法,ユークリッド距離)を実施し,放火動機を分類した。その結果,単一放火に関する7類型が見出された(①怨恨型,②自殺型,③不満の発散型,④犯罪副次型,⑤保険金詐取型,⑥火遊び型,⑦人生悲観型)。続く研究2では,連続放火犯127名の供述における44の名詞を変数として,研究1と同様の分析を行ったところ,連続放火に関する5類型が抽出された(①不満の発散(雑多要因)型:ギャンブルの負けなど,②不満の発散(就業要因)型:仕事や就職活動の失敗など,③火事騒ぎ型,④逆恨み型,⑤犯罪副次型)。加えて,単一放火の中心となる動機が“怨恨”であり,連続放火の動機の中心は“不満の発散”であることを示唆した。
著者
財津 亘
出版者
日本法科学技術学会
雑誌
日本法科学技術学会誌 (ISSN:18801323)
巻号頁・発行日
vol.13, no.2, pp.133-142, 2008 (Released:2008-10-17)
参考文献数
30

The relationship between serial rapist type and the time interval from the first to the second crime was investigated in Study 1. A Bayesian Network (BN) model, included rapists' type derived in Study 1 and the time interval from the first to the second crime, was constructed to predict the time interval till the second crime, and this model was tested in Study 2.   In Study 1, 147 serial rapists were classified according to the severity of the rapists' criminality by using categorical principal components analysis (CatPCA). Results indicated that the interval from the first to the second crime for rapists with a low severity of the criminality (mean 158.0 days) was longer than that for rapists with a high severity of the criminality (mean 82.1 days). In Study 2, A BN model was constructed from the following variables: rapists' characteristic derived in Study 1 (the severity of the criminality), time interval from the first to the second crime, as well as criminal behaviors. The model was tested by using new data, 20 serial rapists committed, as virtual cases. The results of model estimation indicated that the accuracy of predicting the time interval from the first to the second crime, either within 42 days or later, was 80.0%. In detail, if the model predicted that a rapist would commit the second crime in less than 42 days, the accuracy of the prediction was 75.0%, whereas if the model predicted that a rapist would commit the second crime in over 42 days, the accuracy of the prediction was 87.5%.
著者
財津 亘 金 明哲
出版者
日本法科学技術学会
雑誌
日本法科学技術学会誌 (ISSN:18801323)
巻号頁・発行日
pp.678, (Released:2014-10-31)
参考文献数
28
被引用文献数
4

The effectiveness of identifying the author of an illegal document by using text mining was investigated. The suspected writing evaluated in this study was a claim of responsibility written by a 14-year-old boy, which stated that he committed the “Kobe child murders” in 1997. It was compared with control writings including confessions, and an essay that we knew were written by the same boy, as well as with irrelevant materials including various essays written by five junior high school students, and claims of responsibility in four past criminal cases. First, the writings in each document were digitalized and converted to text files. Then, the relative frequencies of bigram of letters, bigram of part-of-speech taggers, sentence lengths of each document, and rate of using Kanji, Hiragana, and Katakana were calculated. Results of sammon multi-dimensional scaling and hierarchical cluster analysis indicated that the text in the suspected writing was arranged identically or similarly to groups of texts in control materials, where they were arranged differently from groups of texts in irrelevant materials. In a separate analysis, the suspected writing was substituted with a document written by a different offender and we conducted the identical procedure described above. Results demonstrated that texts in the suspected writing were in a different form control and irrelevant texts. These results indicated the effectiveness of identifying an author by using text mining when examining forensic documents.
著者
財津 亘 金 明哲
出版者
日本行動計量学会
雑誌
行動計量学 (ISSN:03855481)
巻号頁・発行日
vol.46, no.2, pp.87-95, 2019 (Released:2020-03-31)
参考文献数
25

This study examined the difference between authors and the consistency in each author's writing styles, both were the basis of authorship verification. We analyzed 88 academic papers on psychology written by 22 authors and focused on the rates of “non-content words”, “bigram of parts-of-speech”, “bigram of postpositional particles”, “positioning of commas”, ”words before period”, and “Kanji, Hiragana, and Katakana” in the papers. Next, symmetric Kullback-Leibler divergence distances between the papers were calculated. To examine the author differences in writing styles, using hierarchical Bayesian modeling, we compared the distances between papers written by the same author with those by different authors. Furthermore, to examine author consistency in writing styles, we compared the distances of short durations (under five years) between papers written by the same author with the longer durations (over five years). These results supported the hypothesis that there exist author differences and consistency in writing styles.
著者
財津 亘
出版者
公益社団法人 日本心理学会
雑誌
心理学研究 (ISSN:00215236)
巻号頁・発行日
vol.77, no.6, pp.504-511, 2007-02-25 (Released:2010-07-16)
参考文献数
25

The effect of fear on the retrieval of eyewitness' recognition memory was experimentally investigated. Participants were presented with a thematic series of slides of either neutral or ambiguous stimuli. The ambiguous stimulus consisted of a red liquid flowing on a girl's head. Participants who saw the ambiguous stimulus interpreted the liquid as either red paint, or the girl's blood. Differences in the interpretation of the ambiguous stimulus affected the participants' recognition memory. Participants who interpreted the stimulus as blood displayed inferior memory compared to those who interpreted the stimulus as red paint or those who were shown the neutral stimulus. Moreover, when the participant's interpretation was changed from blood to red paint by an experimenter, the memory for central details was enhanced. These results suggest that fear interferes with retrieval. Participants' memory could be improved by reducing fear during the retrieval stage.
著者
財津 亘 金 明哲
出版者
情報知識学会
雑誌
情報知識学会誌 (ISSN:09171436)
巻号頁・発行日
vol.28, no.3, pp.253-258, 2018-09-30 (Released:2018-10-19)
参考文献数
5

本論文は,電子掲示板の書き込みなどによるサイバー犯罪を想定し,その犯人性立証の一手法として,多変量データ解析を用いた計量的文体分析を提案するもので,実際に犯人性の立証が困難であったと思われるいわゆる「パソコン遠隔操作事件」に関する著者の識別を試みた.分析の際は,疑問文章(「パソコン遠隔操作事件」の9事件の文章),対照文章(「パソコン遠隔操作事件」の真犯人であったK氏が過去に敢行し,自供した通称「のまねこ事件」関連の5つの文章),無関係文章(K氏と性別年代が同じ30代男性10名のブログ文章と異なる4つの事件における文章)を対象に,①非自立語の使用率,②品詞のtrigram,③助詞のbigram,④文字のbigramに着目し,階層的クラスター分析を実施した.その結果によると,「パソコン遠隔操作事件」の一連の文章と「のまねこ事件」の一連の文章は,同一人が記載したものである可能性を示唆した.
著者
財津 亘 金 明哲
出版者
情報知識学会
雑誌
情報知識学会誌 (ISSN:09171436)
巻号頁・発行日
vol.27, no.3, pp.261-274, 2017-09-28 (Released:2017-11-24)
参考文献数
35
被引用文献数
1 2

犯人の早期検挙を目的とした捜査支援手法である犯罪者プロファイリングは,犯罪現場の状況などを統計的もしくは心理学的に分析し,犯人の性別や年齢層,職業などの犯人像を推定するものであるが,印字された文書や電子メールなどしか存在しない事件においては犯罪現場自体がないため,有効な手段が従来はなかった.本研究は,文章情報を基に,犯人像を推定する手法の開発の嚆矢として,まずブログを対象に,ランダムフォレストによる著者の性別推定を試みた.その結果によると,性別推定には,漢字や平仮名,片仮名,名詞の使用率,品詞(動詞・形容詞・助詞・感動詞)や接続助詞「し」,助動詞「なかっ」,読点,文字(代名詞「私」「僕」,小書き文字「っ」「ゃ」) の使用頻度が有効で,1個抜き交差確認法による検証の結果,最高で正解率86.0% (適合率:男性84.6%,女性87.5%) を得た.なお,サポートベクターマシンを用いた検証も合わせて実施したが,正解率は最高で 75.0%(適合率:男性69.2%,女性85.7%)と相対的に低い精度を示した.
著者
財津 亘
出版者
日本行動計量学会
雑誌
行動計量学 (ISSN:03855481)
巻号頁・発行日
vol.47, no.2, pp.199-209, 2020 (Released:2021-04-21)
参考文献数
43

An integrated classification algorithm is a decision-making method that is not limited to a single classifier. It comprises multiple classifiers to maintain a high classification performance for various datasets. This study investigated the feasibility of an integrated classification algorithm for offender profiling. Offender profiling is the analysis of a crime scene using statistical and psychological methods to estimate information such as the age, job, and criminal record of the offender. In this study, the following 12 machine learning algorisms were used: decision tree (C5.0, CART by entropy or Gini), logistic regression analysis (LR), naïve bayes (NB), random forest (RF), bagging, boosting, support vector machine (SVM by radial basis function or polynomial), k-nearest neighbor (KNN), and neural network (NN). The results of the study showed that the classification performances of each algorithm varied for different objective variables of the dataset (e.g., criminal record, age, or job of offenders of residential burglar). However, the majority decisions made by a combination of three classifier algorithms (e.g., decision tree, LR, and NB) showed high classification performance regarding any dataset.
著者
財津 亘 金 明哲
出版者
日本行動計量学会
雑誌
行動計量学 (ISSN:03855481)
巻号頁・発行日
vol.45, no.1, pp.39-47, 2018 (Released:2018-11-03)
参考文献数
23

This study examined the accuracy for author identification by text mining. We conducted 16 analyses (four writing styles × four multivariate analyses) across texts of 100 Bloggers, written by approximately 1,000 characters. Specifically, we conducted (1) principal components analysis, (2) correspondence analysis, (3) multi-dimensional scaling, and (4) hierarchical cluster analysis on each writing style: (1) rate of usage of non-independent words, (2) bigram of parts-of-speech, (3) bigram of postpositional particles, and (4) positioning of commas. We obtained high accuracy: 100% on sensitivity and 95.1% on specificity. Furthermore, the results showed no effects of age and gender against accuracy for author identification.
著者
財津 亘 金 明哲
出版者
日本法科学技術学会
雑誌
日本法科学技術学会誌 (ISSN:18801323)
巻号頁・発行日
vol.20, no.1, pp.1-14, 2015 (Released:2015-02-10)
参考文献数
28
被引用文献数
3

The effectiveness of identifying the author of an illegal document by using text mining was investigated. The suspected writing evaluated in this study was a claim of responsibility written by a 14-year-old boy, which stated that he committed the “Kobe child murders” in 1997. It was compared with control writings including confessions, and an essay that we knew were written by the same boy, as well as with irrelevant materials including various essays written by five junior high school students, and claims of responsibility in four past criminal cases. First, the writings in each document were digitalized and converted to text files. Then, the relative frequencies of bigram of letters, bigram of part-of-speech taggers, sentence lengths of each document, and rate of using Kanji, Hiragana, and Katakana were calculated. Results of sammon multi-dimensional scaling and hierarchical cluster analysis indicated that the text in the suspected writing was arranged identically or similarly to groups of texts in control materials, where they were arranged differently from groups of texts in irrelevant materials. In a separate analysis, the suspected writing was substituted with a document written by a different offender and we conducted the identical procedure described above. Results demonstrated that texts in the suspected writing were in a different form control and irrelevant texts. These results indicated the effectiveness of identifying an author by using text mining when examining forensic documents.
著者
財津 亘
出版者
日本犯罪心理学会
雑誌
犯罪心理学研究 (ISSN:00177547)
巻号頁・発行日
vol.53, no.2, pp.29-41, 2016

<p>本研究は,最近10年間の放火事件を対象に,単一ならびに連続放火犯が述べた動機をテキストマイニングによって分類することを目的とした。研究1では,まず単一放火犯が供述した動機に関する文章(文字列)を対象に名詞を抽出し,放火犯ごとで名詞の出現度数を算出した。これにより作成されたデータセット(253名×67名詞)について,古典的多次元尺度法(重み付きユークリッド距離)による分析を行い,名詞を2次元上に布置した。つづいて,2次元上に布置された名詞の座標データを基に,階層的クラスター分析(Ward法,ユークリッド距離)を実施し,放火動機を分類した。その結果,単一放火に関する7類型が見出された(①怨恨型,②自殺型,③不満の発散型,④犯罪副次型,⑤保険金詐取型,⑥火遊び型,⑦人生悲観型)。続く研究2では,連続放火犯127名の供述における44の名詞を変数として,研究1と同様の分析を行ったところ,連続放火に関する5類型が抽出された(①不満の発散(雑多要因)型:ギャンブルの負けなど,②不満の発散(就業要因)型:仕事や就職活動の失敗など,③火事騒ぎ型,④逆恨み型,⑤犯罪副次型)。加えて,単一放火の中心となる動機が"怨恨"であり,連続放火の動機の中心は"不満の発散"であることを示唆した。</p>
著者
財津 亘
出版者
日本法科学技術学会
雑誌
日本法科学技術学会誌 (ISSN:18801323)
巻号頁・発行日
vol.15, no.2, pp.111-124, 2010 (Released:2010-08-27)
参考文献数
23
被引用文献数
1 1

Serial arsonists (N=125) were differentiated into four groups on the basis of their social independence and the degree of criminality, by using categorical principal components analysis (CATPCA). Results indicated the following. (1) Arsonists with high social independence and high criminality were aged 40 years or more in age, did not live with their parents, and tended to have criminal records for theft. (2) Arsonists with high social independence and low criminality were most frequently employed, high school graduates, married, and had no criminal record. Moreover, there were more female and mental patients in this group compared with other groups. (3) Arsonists with low social independence and high criminality were all male, compulsory education level (including high school dropouts), unmarried, not living with parent(s), and had no criminal record of theft. These arsonists tended to have relatively no criminal record of arson compared with other groups. (4) Arsonists with low social independence and low criminality were the youngest among four groups and were aged between 10 and 30 years, unmarried, lived with their parent(s), and had no criminal record. Results of log-linear analysis indicated that arsonists with high social independence tended to use a car or walk to the crime scene, whereas those with low social independence were inclined to use a bicycle. Moreover, arsonists with high criminality records tended to prepare the medium for arson in advance and drink alcohol before the offence, whereas arsonists with low criminality had a tendency to set fire to the same place repeatedly.
著者
財津 亘
出版者
目白大学
雑誌
目白大学心理学研究 = Mejiro Journal of Psychology (ISSN:13497103)
巻号頁・発行日
no.18, pp.1-11, 2022-03-31

本研究は,連続バラバラ殺人における事件リンク分析の有効性を検討したものである。事件リンク分析とは,犯罪捜査の支援を目的として,統計学や心理学的手法を用いて犯行現場を分析する犯罪者プロファイリングの一種である。本研究では,二つの仮説を通して事件リンク分析の妥当性を検討した。一つ目の仮説は,同一犯人による連続殺人の犯罪行動は事件間で類似するといった「行動の一貫性」仮説であり,もう一つは異なる犯人の場合は連続殺人における犯罪行動は事件間で異なるといった「行動の識別性」仮説である。104名による120事件をオンライン新聞データベースより収集し,28カテゴリ(例.殺害方法や遺体の損壊,被害者の遺棄など)を多重対応分析によって検討した。多重対応分析の結果によると,連続犯の犯罪行動はその連続事件を通してまったく同じあるいは類似していたことから,前者の仮説は部分的に支持された。さらに,多重対応分析によると,次元上において104名がそれぞれ敢行した104事件の犯罪行動はまったくお互い合致することがなかったことから,後者の仮説を支持した。
著者
財津 亘 金 明哲 Wataru Zaitsu Mingzhe Jin ザイツ ワタル キン メイテツ
出版者
同志社大学ハリス理化学研究所
雑誌
同志社大学ハリス理化学研究報告 = The Harris science review of Doshisha University (ISSN:21895937)
巻号頁・発行日
vol.59, no.3, pp.181-188, 2018-10-31

本研究は,文章表現において性別を偽装した場合にともなう文体的特徴の変化について実験による検証を行った.実験では,実験参加者48名(男性24名,女性24名,20代から40代で等しく人数を割り当てることで性別年代を統制)が,ブログから抽出した原文を異性の文章に書き換えるといった課題を行った.分析の結果によると,男性実験参加者が女性に偽って文章を書き換えることで,「漢字」や「助詞(連体化)」,一人称代名詞「僕」「俺」の使用率が減少し,「ひらがな」や「動詞(非自立)」,一人称代名詞「私」の使用率が増加した.他方,女性実験参加者が男性を偽る場合は,「漢字」の使用率が増加する,または一人称代名詞「私」の使用率が減少するなど,男性実験参加者とは反対の変化がみられた.この他,「品詞(名詞,感動詞など)」や「カタカナ」,「読点」,「小書き文字」,「終助詞」,「読点前の文字」の使用率における変化はみられなかった.