著者
井上 寛康 イノウエ ヒロヤス Hiroyasu INOUE
雑誌
大阪産業大学経営論集
巻号頁・発行日
vol.11, no.2, pp.165-176, 2010-02

In social science, it is mostly assumed that there is a typical value for data. This is because normal distribution is supposed. However, it is hardly possible that all data in the natural world as well as human society obey normal distribution. Hence, if people take data for some phenomenon and suppose the data obeys normal distribution, that can lead to a wrong conclusion. In fact, people can have accessed many large scale empirical data in recent years and found those distributions do not obey normal distribution. Some of them obey power-law distribution whose nature has been recently well studied. This paper studied whether the distribution of joint patent applications for an organization obeys power-law distribution or not. The distribution is far from normal distribution and looks like power-law distribution. Hence, it was judged whether the distribution obeys power-law distribution using Clauset's method. The results showed the distribution in 7 and over obeys power-law distribution, and the exponent is 3.03.

言及状況

Twitter (1 users, 1 posts, 0 favorites)

与えられた確率 分布がどの程度にべき分布に近いのかということを定量化し,妥当な閾値によって判別 できる方法が,Clauset らによって最近開発された https://t.co/2gshuffR5p

収集済み URL リスト