Forecasting the FTSE 100 with high-frequency data: A comparison of realized measures (not a question)

Reference page for the FEX submission related to my Dissertation (MSc in Finance & Economics 2010/11)

LINKS

Link to the code on FEX: MATLAB code

Link to the paper on SSRN: Forecasting the FTSE 100 with high-frequency data: A comparison of realized measures

STATS

Some stats from SSRN on my paper (using Trendy: http://www.mathworks.com/matlabcentral/trendy/plots/741)

CONTENT OF SUBMISSION

Brief contents of the submission:

• 20110916_Dis.m
• bp.m
• fltmedian.m
• fltout.m
• fltprice.m
• realized_...m

where the dots after realized_ indicate that there are several realized measures: var, semivar, bpv, tbpv, tripv.

The private folder contains:

• fltret.m
• Modified Bruno Luong's mcolon.m and similar

Other needed submissions

• My regstats2.m
• Paul Mattern's DataTable.m

FUNCTION TODO LIST

• Implement the corrected realized_tbpv.

ANALYSIS TODO LIST

• Add Model Confidence Set stats (from 2010 Hansen, Lunde, Nason - The Model Confidence Set - WP)
• Add economic measures of forecasting performance (such as Sharpe Ratio etc...)

KNOWN ISSUES

• If a day has a whole (missing data) filtret.m does not pad the previous observation as it should under 'CalendarTime', 'Last'. In my analysis the problem occurs only a couple of times which are excluded during the cleaning of the data.

HISTORY

• 01 Oct 2011, 17:08 BST - Submitted to the FEX
• 17 Feb 2012, 13:04 BST - Edited description of this page and added trendy plot.

Feedback of ANY kind is welcome.