tag:blogger.com,1999:blog-3884463987282087684.post3375677441046405265..comments2024-06-06T10:51:18.020-07:00Comments on DataTechNotes: Regression Example with Keras LSTM Networks in RUnknownnoreply@blogger.comBlogger10125tag:blogger.com,1999:blog-3884463987282087684.post-50249394518442684632023-08-27T00:49:26.161-07:002023-08-27T00:49:26.161-07:00Hi, you can not use all data to train the net, sin...Hi, you can not use all data to train the net, since you use it to predict those data in use and absolutely it does perfectly. But your model should be tested with another data set, which operates very badly.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-90444990839924533722019-11-05T20:45:42.658-08:002019-11-05T20:45:42.658-08:00Hi, I tried this method work for time series data ...Hi, I tried this method work for time series data with last 4 year monthly values. The predicted values are vague and i'm not sure of what i did wrong. I also tried by changing the step size but it is also not working out.Can you please help me out with it?Anonymoushttps://www.blogger.com/profile/04509875863044318291noreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-85446824043715501362019-08-26T14:12:40.107-07:002019-08-26T14:12:40.107-07:00Hi, I followed your advices and my model has impro...Hi, I followed your advices and my model has improve, thanks. But (another doubt) I got some steps (number of samples to reach a new period) with differents numbers, at the beginning of the sample, each 4 samples change the period, and at the end, that changes for 5 periods. How should I attack this problem? 2 models? how can be ensembled?<br /><br />Thanks for your time, really.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-23522437314592760362019-07-17T23:22:51.421-07:002019-07-17T23:22:51.421-07:00Good! Your data is too small to evaluate your mode...Good! Your data is too small to evaluate your model and improve the performance. To check the improvement in your model;<br />1) Use bigger data,<br />2) Change the units number,<br />3) Add dense layer,<br />4) Add dropout layer, layer_dropout()<br />5) Change optimizer (rmsprop etc.)DataTechNoteshttps://www.blogger.com/profile/08605431102919145423noreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-36244366247808981122019-07-17T11:40:16.865-07:002019-07-17T11:40:16.865-07:00Thanks, I made an X array with all the predictors ...Thanks, I made an X array with all the predictors and it works. Got a mse = 15.9 (nice) with the default parameters, then I tunned the epochs parameter on the fit and got a better prediction. I´ve been tunnin with epochs and batch_size but I dont know very well how should I change the sequential keras model, (dense and units), I got 37 observations and 19 predictors. Can you give me advices with this tunning? Thanks for your time and post, my model's predictions are great, in fact I could stop now with my results but I want to improve and learn more about this model.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-23872001995921608742019-07-10T17:34:15.435-07:002019-07-10T17:34:15.435-07:00You are welcome! You need to create combined X arr...You are welcome! You need to create combined X array data (contains all features x1, x2, ..) for your training and prediction. It goes like this;<br />x1, x2, y<br />2, 3, 3<br />3, 4, 4<br />2, 4, => 4<br />3, 5, => 5<br />4, 6, => 6<br /><br />Here, each window contains 3 elements of both x1 and x2 series.<br />2, 3,<br />3, 4,<br />2, 4, =>4<br /><br />3, 4,<br />2, 4,<br />3, 5, => 5<br /><br />2, 4,<br />3, 5,<br />4, 6, => 6DataTechNoteshttps://www.blogger.com/profile/08605431102919145423noreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-68899341083751470182019-07-10T07:44:47.581-07:002019-07-10T07:44:47.581-07:00How do I tune this model?How do I tune this model?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-9696282580517159352019-07-09T14:18:26.230-07:002019-07-09T14:18:26.230-07:00Hello, excelent post, Im in a proyect using this a...Hello, excelent post, Im in a proyect using this algorithm and I have one question, if I have more predictors, on the model fit should I use ###fit(x1+x2,y,....) and the predictions ###predict(x1+x2) ??? or am I wrong?<br />Thanks for your help. Great post.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-54719727813521317712019-06-24T21:37:56.739-07:002019-06-24T21:37:56.739-07:00Good point! But, here I did not intend to build a ...Good point! But, here I did not intend to build a perfect predictive model. The purpose of this post is to show a simple, workable example with a random data for beginners. Readers should consider every aspect of the model building when they work with real problems.DataTechNoteshttps://www.blogger.com/profile/08605431102919145423noreply@blogger.comtag:blogger.com,1999:blog-3884463987282087684.post-46588904637198459982019-06-24T13:10:25.603-07:002019-06-24T13:10:25.603-07:00This model only looks good because it probably ove...This model only looks good because it probably overfits the data. You did not include any test/validation data to see if the model generalizes out of the training sample. Additionally, with only 400 data points but almost 80,000 learnable parameters, the memory capacity of the net is likely too large for this task. This means that the net was probably able to memorize the test data's specific input-output mappings, and will thus lack predictive power. Anonymousnoreply@blogger.com