search

kongming01 / MDL

Detecting Product Adoption Intention via Multi-View Deep Learning
4 stars 1 forks source link

[Help Wanted] Waiting for your codes #1

Open mokeeqian opened 2 years ago

mokeeqian commented 2 years ago

I am interesting for you paper within a MDL framework, could you please offer your codes for MDL? Thanks a lot!

kongming01 commented 2 years ago

Hi, please see the file fasttext_cnn_lstm.py for my codes of MDL! Thank you for your interest in my paper.

mokeeqian commented 2 years ago

Hi, please see the file fasttext_cnn_lstm.py for my codes of MDL! Thank you for your interest in my paper.

okey, I will have a try, thank you!

mokeeqian commented 2 years ago

emmmm, the code label_list = list(open(data_location + dataset + "/graph.txt").readlines()) in the funtion load_data11 of data_helpers.py, what does the file graph.txt mean? And where is it? Sinceerely, yours.

kongming01 commented 2 years ago

Sorry, my current codes are too messy, as I focused on testing the research idea. The original file "graph.txt" contains two tweet IDs separated by a Tab space in each line, which means a kind of connection between two tweets.

For the codes you see now, they can actually take both text of tweets and the graph information of tweets as two parts of input, but in the final version of my paper for MDL, I don't use the graph information. Therefore, a simple solution is to create a text file named "graph.txt " without anything in it, then the files fasttext_cnn_lstm.py and data_helpers.py could run well without using the graph information.

Another way is to remove the codes related to the graph information, and this will not impact the effect of MDL as it only uses text information. However, recently I have no time to revise the codes and make them concise.

mokeeqian commented 2 years ago

Sorry, my current codes are too messy, as I focused on testing the research idea. The original file "graph.txt" contains two tweet IDs separated by a Tab space in each line, which means a kind of connection between two tweets.

For the codes you see now, they can actually take both text of tweets and the graph information of tweets as two parts of input, but in the final version of my paper for MDL, I don't use the graph information. Therefore, a simple solution is to create a text file named "graph.txt " without anything in it, then the files fasttext_cnn_lstm.py and data_helpers.py could run well without using the graph information.

Another way is to remove the codes related to the graph information, and this will not impact the effect of MDL as it only uses text information. However, recently I have no time to revise the codes and make them concise.

okey, thanks a lot! hope your studies go well!