and helps prevent erroneous links between stories and entities. In particular, after
filtering by relevance as measured by RavenPack, Hafez (2009a) obtains a 3improve-
ment in correlations between a calculated market sentiment measure and out-of-sample
returns. Both Reuters and RavenPack include measures for article novelty (uniqueness)
which determines repetition among articles and how many similar articles there are for a
particular company. In addition, RavenPack (2010) measures event novelty based on
more than 200 event categories that are automatically detected in the news. This allows
the user to consider not only the first instance of a company event but also to measure
how much media attention it receives.
Several studies also reportstrong seasonality in news flowat hourly, daily and weekly
frequencies (Lo, 2008; Hafez, 2009b; Moniz, Brar, and Davis, 2009). A valuable aspect
of pre-analysis of news data is to identify periods of unexpected news flow levels, from
periods of variation due to seasonality, in order to identify periods where significant
levels of information are flowing into the market. Hafez (2009b) investigates the season-
ality patterns of news arrival. Figures 1.4 and 1.5 show the intraday pattern. He notes
that larger volumes of news flow arrive just before the opening of the European, US, and
Asian trading sessions. On the intra-week level we can see little news flow takes place at
the weekends. During the week, the peak of news flow occurs on Wednesday and
Thursday, while the trough falls on Friday. Lo also notes that the median number of
weekday Reuters news alerts is usually between 1,500 and 2,000, while the median for
the entire weekend drops to around 130.
Thetime of the day when news is releasedhas also been found to be relevant in
understanding the connection between market variables and news. Robertson, Geva,
8 The Handbook of News Analytics in Finance
Figure 1.3.CARs start to respond several days before relevant news is published.