gcooq
embedding
Hello， I want to ask how to produce embedding vector ? Just use word2vec ?
Thanks for your attention, you first split the long trajectory to short trajectory with time interval. Then you can use the word2vec tookite training the datasets
You have defined your loss function, so do you call word2vec directly ? Your embedding vector has been trained in advance and doesn't exist in your code.
I used the tool of word2vec, you can download this toolkit via python command. First pretraining all of the trajectory data by word2vec and saved these vectors, then you can use the sourcecode . I will upload the details and some about datasets soon.
Hello, I want to know how you concatenate the location points within six hours of the original Gowalla data into a track. Can you share the code?
