A doubt about the time series chain

[adityabhandwalkar]

Hello Community,

I have a general query about the time series chain and general motif search:

1. With time series chain , as we advance further into the time series data, patterns do evolve in significant manner and they tend to elaborate the process in terms of change happening over the course of time. While using the allc we compute the unanchored chain, My question is during the process when we opt for the different regions of chains and they have the different colors and of different size , is it possible to label them based on feature of size and categorized them in terms of patterns which are normal , not so normal and abnormal . Secondly while other region in time series is not colored and remains same does this mean that the only colored region seems to have evolution and rest of the data is normal? as shown in figure

2. With motif search a can we rank them as in normal , not so normal and abnormal(try to plot them for e.g in Red , yellow or orange according to their normality) just for the clarification of the extracted motif patterns or it is completely useless to classify the motif within the categories.

@seanlaw Do you have any thoughts/remarks about this ?

[seanlaw]

@seanlaw Do you have any thoughts/remarks about this ?

Sorry, I do not have much experience with time series chains.

[adityabhandwalkar]

Do you have any suggestion on point number 2 , for classifying the motifs?

[NimaSarajpoor]

I think your should take a look at the original paper, and then the tutorials in stumpy/docs. (And, by take a look, I mean... read and understand it)