TAQ data and Eric Zivot's HF Library
On Thursday 15 February 2007 02:12 pm, Panov, Evgeny [CIB-EQTY] wrote:
Did anybody try to look at co-skewnesses for stocks for different sampling frequencies? (i.e. 1-week as opposed to 1-day as opposed to 5-minutes)
I realized that no one responded to this post. We had previously released functions for the co-moments to this list, and CoSkewness is included in the code we released as the PerformanceAnalytics package this week. You could use wavelets (via the package waveslim) or the aggregate function from zoo to construct the series for the various frequencies that you want to analyze. Regards, - Brian
773-459-4973 mobile http://braverock.com/brian/resume-quant.pdf