Hi, I'm a student at NITW,INDIA reimplementing an RNN-based ransomware detection framework
(RansoGuard, Cen et al. 2025) that requires the ORDER of API calls per sample (not just frequency/presence counts).
Your feature-selection step starts from 6.4M features, which suggests the
released dataset in 5_mlran_dataset is a flattened/tabular representation.
Question:
Are the original raw Cuckoo report.json files (with the ordered "calls" list per process) available anywhere — e.g. on Zenodo or another host —
or is the flattened CSV the only released format?I am particularly interested in the sequence of API calls so thought i would ask...
Thanks for putting this dataset together, really useful work.
Hi, I'm a student at NITW,INDIA reimplementing an RNN-based ransomware detection framework
(RansoGuard, Cen et al. 2025) that requires the ORDER of API calls per sample (not just frequency/presence counts).
Your feature-selection step starts from 6.4M features, which suggests the
released dataset in 5_mlran_dataset is a flattened/tabular representation.
Question:
Are the original raw Cuckoo report.json files (with the ordered "calls" list per process) available anywhere — e.g. on Zenodo or another host —
or is the flattened CSV the only released format?I am particularly interested in the sequence of API calls so thought i would ask...
Thanks for putting this dataset together, really useful work.