w5688414 / amnn

AMNN: Attention-based Multimodal Neural Network Model for Hashtag Recommendation

AMNN

AMNN: Attention-based Multimodal Neural Network Model for Hashtag Recommendation

Dataset:

We collect almost 248,166 public microblogs according to selected 97 hashtags of "Top 100" on Instagram. The final collection contains 56861 microblogs which include both text and image, called MultiModal data from Instagram (MM-INS). We filter duplicate hashtags in one sample and drop out those microblogs without texts.

This dataset is a collection of crawled microblogs from Instagram by using Instaloader API, https://instaloader.github.io/. As the raw dataset is too larger to upload all of them, we choose 3 sub-datasets without preprocessing, including "#beach", "#cat", "#dog", and the corresponding sub-datasets with preprocessing that remove those images without texts, including "beach", "cat", "dog". In addition, more samples can be found in Google Driver. Hope the data can be helpful for your research, and we are open for academic cooperation if necessary.

Open access:

• IEEE Dataport: https://ieee-dataport.org/documents/multimodal-dataset-instragram
• Google Driver: https://drive.google.com/drive/folders/1WjKwfMpSv-wBQW5CMXSAgkkwGhOnJawg?usp=sharing

amnn's Issues

Questions about licenses.

Hello. I'd like to use your code for non-profit services. I want to upload the code to my GitHub. What's the license?

Request for the complete MM-INS dataset

I am trying to experiment with the code, however due to unavailability of the complete dataset I am unable to validate my results. Could you please provide a link to the entire MM-INS dataset.