Timeline for LSTM doesnt find finer dependencies than the Random Forest model
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 21, 2018 at 8:58 | comment | added | Meiiso | @Gino_JrDataScientist 1.Of course I test on a validation set. 2.The results are only that good because I've done pretty good feature engineering. My concern here is that I bring the advantages of the two models in one. The question is: how? | |
| May 21, 2018 at 8:46 | answer | added | Gino_JrDataScientist | timeline score: 1 | |
| May 21, 2018 at 8:42 | comment | added | Gino_JrDataScientist | Also: your data might have both a daily dependence e.g. higher call volume in the afternoon) and a weekly dependence (more calls on Monday). Instead of training a model on the whole series of call volume, you might be better off feeding your model one day at a time, that is you cut the time series between 8 am and 5 pm every day. After all, you do not want your model to spend parameters on learning that no calls are made at night or during the weekend. Plus you can add the day of the week as a parameter in the model | |
| May 21, 2018 at 8:39 | comment | added | Gino_JrDataScientist | Is this the training data? Make sure that you are testing your models on an independent sample. A common way is to split the dataset in two, train on one (say time between 0 and 500) and test on the other. Then look at the R2. The reason I'm asking is that I am very surprised to see such good results. | |
| May 20, 2018 at 18:43 | history | undeleted | Meiiso | ||
| May 20, 2018 at 18:43 | history | deleted | Meiiso | via Vote | |
| May 20, 2018 at 18:37 | history | asked | Meiiso | CC BY-SA 4.0 |