- Files
- jaguar1637
- Problem of selecting the best lambda in Hodrick prescott algorythm

Hi All

I met some deep problems regarding the calculation of the variable Lambda in my indicators

It 's sur to the fact, the Hodrick-Prescott (HP) requires setting a value for the smoothing parameter

λ. This value should be chosen based on statistical properties and economic

we want to see fulfilled by the trend and the cycle thus obtained.

• On a statistical level, choose the value of λ is equivalent to selecting from fluctuations

which are the short term and the movements that affect the long term. in

practice, λ is too low affects wrong part of the cycle to the short periodicity

trend leading the latter to be too volatile. On the other hand, a too high λ leads to

overestimate the cyclical component. Choose the parameter λ is therefore to

determine the average length of the cycles of activity. This criterion not select a value too

high for λ.

• On an economic level, choose a highly volatile trend returns to lead a

structural analysis in an economic environment stabilized enough, otherwise

said too influenced by cyclical fluctuations. This criterion not select

too low a value for λ.

So, I am wandering of taking into account the result from MovingMinMax indicator or the fractals dimension, or the matrix correlation dimensio....giving the regular period.