The ACODE dataset has been released!

acode_logo

We are delighted to announce that we share the ACODE dataset with the research community. It consists of text descriptions of 200K Android apps. The apps and their descriptions were collected from official market place (Google Play) and third-party markets. The text descriptions are written in English and Chinese. In addition to the raw text descriptions, the dataset includes manually labeled text descriptions where a label indicates whether a text description refers to use of a particular permission or not. Furthermore, we also share key extracted results (keywords for classifying text descriptions) so that other researchers can reproduce our results. More details as well as the way to download the data are available from the ACODE project page. Please feel free to drop us an e-mail message if you want to try out. Enjoy!