Welcome to the Regime Shift Detection Project
After closing the full-featured Climate Logic website, I received a number of requests to post the latest version of the regime shift detection software, often referred to as (RSD), or sequential t-test analysis of regime shifts (STARS). The RSD is convenient tool that automatically detects regime shifts in a time series, in both the mean and variance. It has a number of advantages over other existing methods: better performance at the end of time series, indicates a possibility of a shift in real time, can be tuned up to detect shifts of different scales, works well in the case of a moderate linear trend, no need to calculate anomalies prior to applying the method, etc. It can also handle time series with autocorrelation, thus separating regimes due to persistence from the true regimes with different statistics.
The earlier version of the RSD (v. 3.2), which is hosted at the Bering Climate website, has some bugs affecting the calculation of regime shift in variance. Those bugs are fixed in v. 3.4 available here for download. Please keep in mind that, since the software is free, it is not supported. However, if you find a bug, I'd appreciate if you send me a note.
As for the other features of the old Climate Logic website (climate news, seasonal forecasts and reviews, decadal variability and trend analyses, weather derivatives, etc. ), I may add them it in the future, if I have enough time. Currently, I am 100% focused on monthly temperature forecasts for the United States, using weather derivatives trading as a verification tool.
Regards,
Sergei Rodionov
Climate Logic, LLC .
After closing the full-featured Climate Logic website, I received a number of requests to post the latest version of the regime shift detection software, often referred to as (RSD), or sequential t-test analysis of regime shifts (STARS). The RSD is convenient tool that automatically detects regime shifts in a time series, in both the mean and variance. It has a number of advantages over other existing methods: better performance at the end of time series, indicates a possibility of a shift in real time, can be tuned up to detect shifts of different scales, works well in the case of a moderate linear trend, no need to calculate anomalies prior to applying the method, etc. It can also handle time series with autocorrelation, thus separating regimes due to persistence from the true regimes with different statistics.
The earlier version of the RSD (v. 3.2), which is hosted at the Bering Climate website, has some bugs affecting the calculation of regime shift in variance. Those bugs are fixed in v. 3.4 available here for download. Please keep in mind that, since the software is free, it is not supported. However, if you find a bug, I'd appreciate if you send me a note.
As for the other features of the old Climate Logic website (climate news, seasonal forecasts and reviews, decadal variability and trend analyses, weather derivatives, etc. ), I may add them it in the future, if I have enough time. Currently, I am 100% focused on monthly temperature forecasts for the United States, using weather derivatives trading as a verification tool.
Regards,
Sergei Rodionov
Climate Logic, LLC .