TemporalModelling&Forecasting
0 likes275 views
This document discusses new temporal (daily, weekly, monthly) modeling and forecasting in the FX market. It presents models for high, low, open and close prices across different timeframes. For monthly, weekly and daily data on XAUUSD, XAGUSD and WTI, the document shows modeling of high and low prices using local regression and wavelets with a goodness of fit between observed and modeled data of 99.99%. Forecasting examples are provided for weekly and monthly high/low prices of XAGUSD and monthly high prices of WTI and XAUUSD through 2014-2015. The conclusion states the temporal modeling is very close to observed data and useful for FX trading.
TemporalModelling&Forecasting
- 1. New Temporal (Daily, Weekly and Monthly) Modelling and Forecasting in the FX Market Shiquan REN, PhD ctusren@outlook.com https://au.linkedin.com/in/ctusren April 8, 2015
- 2. Contents • Background • Temporal (Daily, Weekly and Monthly) Data • Temporal (Daily, Weekly and Monthly) Modelling • Forecasting/Prediction • Conclusion
- 3. Background • There are so many temporal data - monthly, weekly and daily (high/low) data in the forex market • Objective: I try to build and develop a series of temporal models to forecast the monthly, weekly and daily (high/low) data distribution based on the relationship of temporal data such as high, low, open and close data
- 4. High, Low, Open and Close Data • DAILY • XAUUSD (1/10/2006 – 7/04/2015) • XAGUSD (1/10/2006 – 7/04/2015) • WTI (1/03/2008 – 7/04/2015) • WEEKLY • XAUUSD (20/04/1998 – 3/04/2015) • XAGUSD (30/12/1996 – 3/04/2015) • WTI (18/02/2008 – 3/04/2015) • MONTHLY • XAUUSD (1993/02 –2015/03) • XAGUSD (1997/01 –2015/03) • WTI (2008/02 –2015/03)
- 5. Temporal Modelling • Daily, Weekly and Monthly Modelling of High/Low Prices • Local regression and Leave-one-out samples without overfitting (Ren et al, 2010; Ren and Kingsford, 2014) • Voronor-based lifting wavelets (Jansen et al, 2009) • New models and algorithms
- 6. Monthly Modelling of XAUUSD High Prices The goodness of fit between the observed and modelled data is 99.99%
- 7. Monthly Modelling of WTI Low Prices The goodness of fit between the observed and modelled data is 99.99%
- 8. Weekly and Monthly Forecasting of XAGUSD High/Low Prices 2015/04/01~30 2015/04/06~10 model quantile AXI FXCM AXI FXCM high upper 19.670 18.531 25.773 21.348 high 80% 18.619 18.035 21.034 20.240 high 61.8% 18.346 17.890 20.269 19.164 high median 18.268 17.604 19.288 18.282 high 38.2% 18.190 17.318 18.307 17.037 high 20% 18.112 16.952 17.812 16.901 high lower 18.007 16.630 16.716 16.697 low upper 16.614 16.630 16.716 16.697 low 80% 16.614 16.630 13.527 14.108 low 61.8% 16.614 16.630 13.111 11.791 low median 16.614 16.630 12.446 10.852 low 38.2% 16.614 15.424 11.781 9.913 low 20% 16.365 14.835 11.297 8.870 low lower 15.579 14.609 10.540 8.244
- 9. Monthly Forecasting of WTI High Prices in June 2014 The highest price was 107.5 on 2014/06/15 model quantile 2014/06 high upper 107.5463 high 80% 106.8884 high 61.8% 106.4498 high median 106.0296 high 38.2% 105.6094 high 20% 105.4324 high lower 105.2237 low upper 94.60887 low 80% 93.10167 low 61.8% 92.56926 low median 91.48316 low 38.2% 90.39706 low 20% 89.22345 low lower 88.20962
- 10. Monthly Forecasting of XAUUSD High/Low Prices in November 2014 The lowest price was 1131.67 on 2014/11/07, the higher price was 1178.70 on 2014/11/07~10, and then the lower was 1147 or so among 2014/11/10~14, and the highest price was 1207.75 on 2014/11/21 model quantile 2014/11 high upper 1306.591 high 80% 1220.265 high 61.8% 1207.159 high median 1177.635 high 38.2% 1172.49 high 20% 1172.49 high lower 1172.49 low upper 1172.49 low 80% 1148.276 low 61.8% 1131.205 low median 1107.408 low 38.2% 1083.612 low 20% 1072.59 low lower 1020.988
- 11. Conclusion • My new temporal modelling, which is the unique and best, is very close to the observed data. The goodness of fit between the observed and modelled temporal high/low data is 99.99% • My own forecasting/prediction of temporal (daily, weekly and monthly) high/low prices is useful and powerful in the FX trading.