Hi Every body, In order to deal with nonstationary problem in time series, may be firstly clustering algorithms are used to partition time series .Then another algorithm is used to predict future value based on segmented data in the second phase. Using clustering algorithms , the "time structure and arrangement" of time series is confused. We have some partitions including data unrelated to the time at hand. A question which arises here is that: lossing the time arrangement of time series is not a new problem? can we forecast the future based on segmented confused clusters? What am i missing? Have a nice Amir --------------------------------- [[alternative HTML version deleted]]
Did you thought of clustering with restriction that in each cluster, time periods must all be "connected"? ----- Original Message ----- From: "Amir Safari" <amir36060 at yahoo.de> To: <R-help at stat.math.ethz.ch> Sent: Friday, May 13, 2005 11:37 PM Subject: [R] clustering> Hi Every body, > In order to deal with nonstationary problem in time series, may be > firstly clustering algorithms are used to partition time series .Then > another algorithm is used to predict future value based on segmented data > in the second phase. Using clustering algorithms , the "time structure and > arrangement" of time series is confused. We have some partitions including > data unrelated to the time at hand. A question which arises here is that: > lossing the time arrangement of time series is not a new problem? can we > forecast the future based on segmented confused clusters? What am i > missing? > > Have a nice > Amir > > > --------------------------------- > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! > R-project.org/posting-guide.html > >