07-31-2017 11:24 AM
First of all, thank you for your time for my question.
I am currently analyzing the video viewing history data of many people.
However, this data does not have a user's score, so I need to create a score.
Here is my data: This gubun variable indicates that the user is in the following state:
USER GUBUN MOVIE
1 F A
1 B A
1 W A
1 W B
1 B B
1 W B
2 W C
2 W C
2 B D
2 B A
3 W E
3 B B
3 F E
3 B B
F is favorite, B is buy, W is watch
The data listing means that the action has taken place in that order.
I want to score the continuous behavior that users have been doing.
For example, a continuous F-B data set appears. In other words, I would like to treat it as a single action and assign it to a new variable if I made a purchase after clicking the Like button. I do not know how to evaluate continuous data.
user F-B-W F-B W-B-W W-W B . W-F .... (many combination possible )
1 1 1 1 0 2 0 ....
2 0 0 0 1 2 0
I just want to know how to count sequence in ordered data. please give me some advice.
Thanks in advance for your help.
07-31-2017 12:19 PM
A few items that you will need to either confirm or clarify ...
If a user has no "B" observations, then it seems you want no output for that user.
The sequences are determined for each User / Movie combination.
If a user's "B" observation occurs on only one movie, you still want the combinations for the other movies that don't have a "B".
What is the "Like button"? Is there a variable in the data set that indicates "like"?
07-31-2017 12:30 PM
I want to make output even there is no B observations. for example , W-W also important ,
for modeling, i count many various sequences.
it was my mistake, i didn't write that like button means gubun =F ,
07-31-2017 12:52 PM
OK, one question still unanswered. What happens when ID=3 and the movie values alternate? Are the sequences W-F and B-B? Or do the sequences get measured by ignoring the MOVIE and just looking at GUBUN?