[Eventstudies-commits] r264 - pkg/man

[noreply at r-forge.r-project.org]

Author: vikram
Date: 2014-03-30 09:31:10 +0200 (Sun, 30 Mar 2014)
New Revision: 264

Modified:
   pkg/man/ees.Rd
   pkg/man/eesPlot.Rd
Log:
Added details for extreme event analysis function

Modified: pkg/man/ees.Rd
===================================================================
--- pkg/man/ees.Rd	2014-03-29 18:00:14 UTC (rev 263)
+++ pkg/man/ees.Rd	2014-03-30 07:31:10 UTC (rev 264)
@@ -2,12 +2,10 @@
 \alias{ees}
 
 \title{
-Extreme events study analysis for a univariate time series
+Extreme events study analysis: Summary statistics
 }
 
-\description{This function is used to identify tail events on a
-  univariate time series. It generates summary statistics for clustered and
-  unclustered tail events.}
+\description{ This function generates summary statistics table from Patnaik, Shah and Singh (2013), using a univariate time series.}
 
 \usage{
 ees(input, prob.value)
@@ -22,6 +20,17 @@
    distribution for the tail event.}
 }
 
+\details{ Patnaik, Shah and Singh (2013) uses modified event study methodology in the paper, which is used to understand the impact of extreme tail events on a given variable (foreign investment, stock market returns in this case). Event dates are defined as those on which extreme values of returns or flows are observed. The argument \sQuote{prob.value} defines the extreme tail values on both side of the distribution as event dates.  For instance, if \sQuote{prob.value} is 2.5 then it will consider 2.5 percent tail values on both side of the empirical distribution of the variable as event dates. The paper further differentiates between clustered events (consecutive extreme events) and unclustered events which is described in detail in summary tables.
+This function helps to understand the various feature of the event dates by providing following summary statistics:
+   \itemize{
+    \item \dQuote{Distribution of extreme events}:  Number of events in clustered, unclustered and events used and discarded from the analysis
+    \item \dQuote{Run length distribution}: Table shows number of clustered events with a particular run length
+    \item \dQuote{Quantile values}: Quantile wise distribution of extreme events
+    \item \dQuote{Yearly distribution}: Year wise distribution of extreme events
+  }
+  Please refer to Patnaik, Shah and Singh (2013) for more detailed explanation.
+}
+
 \value{ A \code{list} object containing:
 
   \item{data.summary}{a \sQuote{data.frame} containing summary of

Modified: pkg/man/eesPlot.Rd
===================================================================
--- pkg/man/eesPlot.Rd	2014-03-29 18:00:14 UTC (rev 263)
+++ pkg/man/eesPlot.Rd	2014-03-30 07:31:10 UTC (rev 264)
@@ -2,12 +2,11 @@
 \alias{eesPlot}
 
 \title{
-Plotting clustered and unclustered extreme events retuns.
+Extreme event study analysis: Plotting clustered and unclustered extreme events returns
 }
 
-\description{The function creates an event study plot that treats all
-  clustered events as one event. It plots both the lower and upper tail
-  events.}
+\description{ This function uses modified event study methodology from Patnaik, Shah and Singh (2013) to plot the impact of event series on the response series. 
+}
 
 \usage{
 eesPlot(z,
@@ -40,10 +39,11 @@
     percentage terms).}
 }
 
-\details{
-  The events are defined based on the cut off probability values
-  provided (in percentage terms).This function replicates result of
-  Patnaik, Shah, and Singh (2013).
+\details{Patnaik, Shah, and Singh (2013) use modified event study methodology to draw the event study plot. The function draws the event dates from the tail of the event series. Extreme events are then formatted by fusing the consecutive events on same side of the tail into one event, further fusing the response series on a clustered event and removing the mixed clusters (consecutive events from left and right tails). After formatting the events, response series is converted to event time frame and mean response is constructed for all event dates. The function also estimates confidence interval using bootstrap inference strategy.  
+
+The paper defines extreme tail events as event dates and the argument \sQuote{prob.value} is used for the same. The function using  the argument \sQuote{event.series.name} picks a column from the data object \sQuote{z}, which is used to extract extreme event dates. While, argument \sQuote{response.series.name} gets the response series on which impact of the extreme event dates is to be analysed. 
+
+Please refer to the paper Patnaik, Shah and Singh (2013) for detailed explanation.
 }
 
 \value{A plot of the response series with lower and upper tail events