|
Post by wangshuaihit625 on Aug 5, 2015 21:37:09 GMT
Hello, I have two questions to ask for help. First, as you know, we have 25 sets of the testing plant data. While this competition inquired us to upload 10 sets of the testing plant data. How do we chose 10 sets of the testing plant data in these 25 sets of the testing plant data(41-43, 45-49, 53-54, and 56-60)? Second, how many predicting steps is better? For example plant 41, there are 1935 current steps. If we predict the next 100 steps, is that OK? Thank you very much. I am looking forward to your reply.
|
|
|
Post by nwilliard on Aug 6, 2015 16:04:44 GMT
You need to predict the missing faults that are in the "plant_XXc.csv" files. There are a total of 15 of these files. You need to submit the csv files with your predicted start time, end time, and fault codes given.
I do not understand your second question...you will want to try to identify all the missing faults to get the best score.
|
|
|
Post by wangshuaihit625 on Aug 7, 2015 14:55:45 GMT
Hello, The second question means that there are 1935 steps as the plant_41c. The last row shows 1935; 2012/11/8 7:14; 6. So I want to know how many steps do we predict? Predicting next 100 steps is OK? The predicting number of steps is from 1935 to 2035.
|
|
|
Post by nwilliard on Aug 10, 2015 1:08:29 GMT
|
|