Join two example sets by relative time value
Thrhw31
New Altair Community Member
Hey,
I'm using Rapidminer for mining some time series data. My imported example set contains several regular attributes and one ID-Attribute called "time". Now I have two of those sets and with different frequencies and precisions.
example set 1:
example set 2
I want to join this sets without getting many missing values. Not like this one:
It would be nice if RM could link the time-ID with the higher precision to the lower precision (closest match, like 0.31 to 0.3). I cant simple round time, because I dont know the precision of my imported data. Perhaps I can use a groovy/java code-snipped, but unfortunately I have no experience with java programming. Every kind of help would be nice
greetings from germany
Thilo
EDIT:
My solution (for now) is
- round the attribute time (two decimal places) of all example sets
- use join to merge the sets
Problem occur if my example set have more than 2 relevant decimal places, like 0.111, 0.112 , 0.113 ...
That's why I'm searching for an alternative solution
I'm using Rapidminer for mining some time series data. My imported example set contains several regular attributes and one ID-Attribute called "time". Now I have two of those sets and with different frequencies and precisions.
example set 1:
time | Attribute X |
0.3 | 3.1 |
0.6 | 2.8 |
0.9 | 2.5 |
1.2 | 2.7 |
... |
time | Attribute Y |
0.11 | 250 |
0.21 | 251 |
0.31 | 248 |
0.41 | 255 |
0.51 | 248 |
0.61 | 253 |
... |
time | Attribute x | Attribute Y |
0.11 | ? | 250 |
0.21 | ? | 251 |
0.3 | 3.1 | ? |
0.31 | ? | 248 |
0.41 | ? | 255 |
0.51 | ? | 248 |
0.6 | 2.8 | ? |
0.61 | ? | 253 |
... |
greetings from germany
Thilo
EDIT:
My solution (for now) is
- round the attribute time (two decimal places) of all example sets
- use join to merge the sets
Problem occur if my example set have more than 2 relevant decimal places, like 0.111, 0.112 , 0.113 ...
That's why I'm searching for an alternative solution
0
Answers
-
It is an old query, but it shows up quite high on Google when searching for a solution to this problem, hence the answer.
One way to join timestamps that have different values and frequencies can be found here:
0 -
I believe that RapidMiner staffer @bhupendra_patil just provided an elegant solution process that will effectively do this for you in a similar thread, found here:
0