I need a C# library which analyse a time serie and run a bunch of checks when values are added to the time serie. A time serie in this case is generic and can be a list of days, months, weeks etc. together with a numeric value (use double as type).
First, I want a class which analyse the time serie with a moving average and regression analysis, preferable with a seasonal index.
When a new value (or a collection of new values) is added to the serie, we shall do a bunch of condition checks.
List of checks:
- Trend (period length as parameter, output: procent)
- Trend broken (true or false)
- Extreme deviation/suddenly dropped by using MSE Min Square Error (MSE maximum value as parameter, output; true or false)
- Has been stable for x points (x as parameter, output: true or false)
- All-time high (true or false)
- All-time low (true or false)
- No data - new value is zero (true or false)
You need to store values in a cache (such as current ”all-time high" value), to more...