023-05

第23回研究会

Supervised LDAモデルによるニューステキストを用いたマクロ経済不確実性指数の構築†

著者†

余野京登, 和泉潔, 坂地泰紀, 島田尚, 松島裕康(東京大学)

概要†

For financial market participants, uncertainty of macro economic events have crucial impact when they make decision to buy or sell financial assets because macro economic event influence the asset price. In this study, we aim to build a model to measure macro economic uncertainty from news text. We proposed extended topic model which using not only news text data but also numeric data as a supervised signal for each news articles. We constructed four macro economic uncertainty indexes by our proposed model. Each indexes matches with historical macro economic events and the correlation are higher with the volatility of market index related to the uncertainty index.

キーワード†

LDA, トピックモデル, テキストマイニング, 不確実性指数

論文†

(10月9日以降に公表いたします)