apache spark - predicting next event from averaging sequences -

i pretty new in ml facing difficulties realizing how use spark machine learning libraries time series data reflect sequence of events.

i have table contains info:

stepn#, element_id, session_id

where step n# sequence in each element appears, element_id element has been clicked , session_id in user session happened.

it consists of multiple sessions , multiple element-sequence per session. i.e. 1 session contain multiple lines of elements. each session have same starting , ending point.

my objective train model use element sequences observed predict next element clicked. meaning need predict next event given previous events.

(in other words need average users click behavior specific workflow model able predict next most-relevant click based on average)

from papers , examples find online understand makes sense when there single sequence of events meant used input training model.

in case though, have multiple sessions/instances of events (starting @ same point) , train averaging model. find bit challenging though understand how approached using example hmm in spark. there practical example or tutorial covers case?

thank spending time read post. ideas appreciated!

this can solve frequent pattern mining. check this: https://spark.apache.org/docs/latest/mllib-frequent-pattern-mining.html

in situation, can find frequent items occurred together. in first step teach model frequent, prediction step, model can see events , can predict common events event