Covering the period January 2005 through December 2008, the research is conducted on the average news flow in a set of pre-specified news aggregation windows ranging intra-day from one- to sixty-minutes; while intra-week and intra-month, a daily aggregation window is applied. Finally, intra-year we use a monthly aggregation window.
The results seem to show a strong degree of seasonality, especially on the intra-day, intra-week, and intra-year horizons with median sector-correlations above 73%, 99%, and 98%, respectively.
On an intra-month level, the sector-correlation is somewhat lower with a median of 40%. Still, 25% of the cross-sectional correlations are above 62%.
Conducting a cross-sectional linear regression analysis on daily news flow based on a set of seasonal dummy variables, we arrive at mean adjusted R-squared of 91.57%.
In order to apply news analytics efficiently in quantitative trading models, it is necessary to consider certain adjustments or normalizations of time series representing news flow. Such normalizations may be relevant as particular seasonality patterns characterize the data.
To identify those times at which news flow is especially relevant to the market, it may be necessary to distinguish true bursts of positive or negative information from mere seasonal peaks in volume. Having prior knowledge of such seasonalities, allows for proper adjustments before conducting data analysis, and thus prevent a wrong interpretation for instance of the impact of increased news flow on volatility and trading volume predictions.
In this paper, the focus will be on the intra-day, intra-week, intra-month, and intrayear news flow seasonality patterns, which are captured by looking at a set of pre-specified news aggregation windows ranging intra-day from one- to sixty-minutes. In addition, for the intra-week and intra-month time horizons, a daily aggregation window is constructed. Finally, a monthly aggregation window is applied for the intra-year analysis.
In Section 2, we introduce our methodology for seasonality detection, which includes an initial seasonality factor identification process followed by a conditional linear regression analysis. In Section 3, we present the empirical results based on historical news flow, while in Section 4, we present our conclusion.
2. Seasonality Detection
Seasonality patterns are often of a magnitude such that they mask other more interesting characteristics of the data. As an example, if each month has a different seasonal tendency toward high or low values, it can be difficult to detect the general direction of a time series’ recent monthly movement (increase, decrease, turning point, no change, consistency with another benchmark indicator, etc.).
Seasonal adjustment produces data in which the values of neighboring minutes, days, weeks, or months are usually easier to compare. Many data users prefer seasonally adjusted data because they want to see those characteristics that seasonal movements tend to mask, especially changes in the direction of the series . That is, performing a seasonality adjustment makes it possible to focus on potential trend and cyclical components in news flow, rather than on the seasonality patterns themselves. Such distinctions are important if working from the premise that the market is driven, not by the seasonal pattern itself, but by unexpected news flow...
Click here to download the PDF and continue reading the "Detection of Seasonality Patterns in Equity News Flows" White PaperRequest White Paper