underdoc-wang
Problems with data dimensions
Hello, the author. Excuse me. I want to ask about Data_ Container_ OD. Load in py_ What do the dimensions read from data represent? Why do I need to use reshape (- 1,47,47)?
Hello. Thanks for your interest in this repo. The shape of data actually represents OD (origin-destination) matrix - essentially (number-of-samples, number-of-origins, number-of-destinations). Supposing we have 47 regions, then the shape of OD matrix should be (47, 47). Hope I solve your problem.
Thank you very much for your reply, it answered my question, and I am very grateful for it！wish you a happy life!
78339239 @.***
------------------ 原始邮件 ------------------ 发件人: "underdoc-wang/MPGCN" @.>; 发送时间: 2022年10月16日(星期天) 凌晨0:10 @.>; @.@.>; 主题: Re: [underdoc-wang/MPGCN] Problems with data dimensions (Issue #2)
Hello. Thanks for your interest in this repo. The shape of data actually represents OD (origin-destination) matrix - essentially (number-of-samples, number-of-origins, number-of-destinations). Supposing we have 47 regions, then the shape of OD matrix should be (47, 47). Hope I solve your problem.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Hello, dear author. I still encountered some problems in the process of model debugging. Our input matrix is (days, number of origins, number of destinations, flow information). For example, (24，238，238, 24) represents 24 days, 238 origins, 238 destinations, and 24 hours of traffic information from origins to destinations , Reshape is (576，238，238）, but the data in the model=[- len (OD_days):,:,:, np. newaxis]. Converting it to a new form leads to construction_ dyn_ G generates NaN, do you want to ask whether this problem is caused by the model input problem? Thank you for very much.
78339239 @.***
------------------ 原始邮件 ------------------ 发件人: "underdoc-wang/MPGCN" @.>; 发送时间: 2022年10月16日(星期天) 凌晨0:10 @.>; @.@.>; 主题: Re: [underdoc-wang/MPGCN] Problems with data dimensions (Issue #2)
Hello. Thanks for your interest in this repo. The shape of data actually represents OD (origin-destination) matrix - essentially (number-of-samples, number-of-origins, number-of-destinations). Supposing we have 47 regions, then the shape of OD matrix should be (47, 47). Hope I solve your problem.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Hi again. You are right. I forgot to mention that here we expand the last dimension (to 1) to fit the input for the 2D graph convolution operation. Thus, you should get (576, 238, 238, 1) as the model input.
