I am test the analysis of Fourier Series in S&P 500 Index some times I have very good results in forecast. The prediction at future is short in time and It´s bad predictor when there ispolitic effect on markets. I fixed two model, exponencial with tend to up and damped:
New tendecy forecast by times series for this week:
TLPV
⋅
Today is the fifth day with very good forescast by time series, yesterday almost in 100% the forescast, the forescast started to be good for low prices but is not trust for top price for the actual tend.
TLPV
⋅
Third day, the forecast with times series used model damped exponential model, is good, but there is sekwed to low extrems.
TLPV
⋅
Update forecast by times series for S&P 500 Index, the model is multiplicative in version exponential damped with 1500 lags, the arrows are variations of prices.
TLPV
⋅
The confidence intervalo is more or less six units since line.
TLPV
⋅
Yes, I am not expert on the stock market, but Ihave two year or more with the study of time series on S&P 500 Index, and I have good forescast fixed model in short time. My analysis was made with sofware Statistical ver 7, the data was obteined the Yahoo Finances, first I search about the lags with technical periodograma, I find in 400, next I observed the tend is more or less exponential,Next I made fixes to find less SMA (Squater Median Average from Error). I found two models with less error, The firts model was exponencial, mulplicative, with 1500 lags. The secound was exponencial and damped tend, multiplicative with 1500 lags.
Jim_MM
⋅
Hey. Can you elaborate a bit on your analysis ? Just being curious but I can't read anything out of your chart, not an expert at fourier series for stock markets