Distillation of News Flow Into Analysis of Stock Reactions
从专业平台、博客和股票论坛的混合文本中提取情绪变量，分析其对股票波动率、成交量和收益的影响，发现情绪效应具有不对称性、关注度特异性和行业特异性。
The gargantuan plethora of opinions, facts and tweets on financial business\noffers the opportunity to test and analyze the influence of such text sources\non future directions of stocks. It also creates though the necessity to distill\nvia statistical technology the informative elements of this prodigious and\nindeed colossal data source. Using mixed text sources from professional\nplatforms, blog fora and stock message boards we distill via different lexica\nsentiment variables. These are employed for an analysis of stock reactions:\nvolatility, volume and returns. An increased sentiment, especially for those\nwith negative prospection, will influence volatility as well as volume. This\ninfluence is contingent on the lexical projection and different across Global\nIndustry Classification Standard (GICS) sectors. Based on review articles on\n100 S&P 500 constituents for the period of October 20, 2009, to October 13,\n2014, we project into BL, MPQA, LM lexica and use the distilled sentiment\nvariables to forecast individual stock indicators in a panel context.\nExploiting different lexical projections to test different stock reaction\nindicators we aim at answering the following research questions: (i) Are the\nlexica consistent in their analytic ability? (ii) To which degree is there an\nasymmetric response given the sentiment scales (positive v.s. negative)? (iii)\nAre the news of high attention firms diffusing faster and result in more timely\nand efficient stock reaction? (iv) Is there a sector-specific reaction from the\ndistilled sentiment measures? We find there is significant incremental\ninformation in the distilled news flow and the sentiment effect is\ncharacterized as an asymmetric, attention-specific and sector-specific response\nof stock reactions.\n