知識共有システムを利用したテキストマイニング手法: 技術経営戦略の立案のために

摘要

Due to globalization, equivocal knowledge including customer needs/various economic affairs in addition to R&D activities is becoming important in order for Japanese companies to design technology strategy from mid- and long-term viewpoint. As one of the techniques for that, statistical analysis of the digitized text data attracts attention now. This is called as text mining. However, conventional text mining techniques remained in acquiring well known knowledge in many cases because analysis is limited in words with much frequency of appearance. This paper proposes to build "an autonomous and distributed knowledge sharing database with write-in space", where "flag", "comments", "link", "words sets" and so on are written in each section of the company. Thus, added values are given to original text data. This is expected to play an important role in decision makings related to technology strategy such as "selection and concentration".%グローバル化の進んだ今日の技術戦略の立案のためには，単に技術に関する知識にとどまることなく，顧客ニーズや経済情勢なども含めた多義的な知識が必要になる．そのための手法の一つとして，テキストマイニングが注目されている．しかし，従来のテキストマイニングは出現頻度の多い単語を多変量解析等の手法で解析するので，その結果は既知の知識にとどまる場合が多かった．本稿は，特許公報の事例をとりあげることにより，企業内に形態素に加えて文章を単位としたマイニングをおこなうことを可能とするような「書き込み空間を持つ自律分散型知識共有システム」を構築し，コメントや単語セット等を企業の各部署で自由に書き込むことによって，事業のあるべき方向性，必要な技術リソース等々の知識を得られることを述べる．