@techreport{oai:ipsj.ixsq.nii.ac.jp:00212219, author = {仲光, 純 and 福田, 悟志 and 難波, 英嗣 and Jun, Nakamitsu and Satoshi, Fukuda and Hidetsugu, Nanba}, issue = {4}, month = {Jul}, note = {特許庁では,技術の発展が期待される分野や社会的に注目されている分野を対象に,特許出願技術動向調査を実施しており,これまでに 250 件以上の報告書と論文・特許リストが公表されている.しかし,これらの報告書は,時間の経過と共に古くなってしまう.そこで,特許出願技術動向調査報告書の自動更新を検討する.本研究では,報告書の自動更新に向けた第一歩として,米国特許の構造を分析することを目的とする.「効果」,「課題」,「手段」といった構造タグが付与された米国特許 4,410 件のデータセットを,構造タグが付与された日本国特許と,それに対応する米国特許(パテントファミリー)を用いて自動作成した.この手法の有効性を確認するために,いくつかの実験を行った.実験の結果,F 値 0.11 が得られた., The Japan Patent Office carries out the patent application technical trend survey for fields in which the development of technologies is expected, or fields to which social attention is being paid, and more than 250 reports and their lists of research papers and patents have been released up to now. Unfortunately, these reports will become obsolete as time goes on. Therefore, we investigate automatic update of patent application technical trend survey reports. In this study, we aim to analyze the structure of U.S. patents as a first step toward the automatic update of the reports. We automatically created a dataset consisting of 4,410 U.S. patents with structure tags, such as “effect,” “task,” and “method,” by using Japanese patents, in which structure tags were manually assigned, and their corresponding U.S. patents (patent families). To confirm the effectiveness of our methods, we conducted some experiments. From the experimental results, we obtained F-measure of 0.11.}, title = {パテントファミリーを用いた米国特許の構造解析および技術動向マップの自動作成}, year = {2021} }