Address – The Bitcoin address
\nFeature 1 – 6 - Different measurements of the Bitcoin transaction graph. They have been anonymized.
\nLabel – The class of ransomware family. There are 8 classes in total, so the label ∈ [0,7].\n\nThe training set should be used to build your machine learning models. For the training set, we provide the correct outcome (also known as the “ground truth”) which is the ransomware family class. Your model will be based on the features in the given training dataset.
\nThe test set should be used to see how well your model performs on unseen data. For the test set, we do not provide the ground truth. It is your job to predict these outcomes. For each row in the test set, use the model you trained to predict the ransomware family the Bitcoin address is from.
\nThe sample_submission file is provided to the contestants as an example submission to see how the submission works in Kaggle.\n\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Description", "content": "This is a Kaggle competition set up for Question 1 of CodeBoss 2020. A maximum number of 10 submissions are allowed for each team.\n\n**The Challenge**\n\nGiven Bitcoin transaction graph data, predict which ransomware family a Bitcoin address belongs to.\n\n**The Story**\n\nRansomware attacks encrypts sensitive information to exchange for a ransom. Attacks have been shown by studies to have caused significant economic losses and the usage of cryptocurrencies like Bitcoin have made it easier for ransomware developers to attack.\n\nBitcoin transaction data are permanently recorded and publicly available. However, current methods to detect ransomware relies only on a couple of heuristics. The idea of exploring data modelling techniques is not widespread as of now.\n\nThis challenge proposes to use data modelling techniques for a data driven method to detect new Bitcoin address from a ransomware family.\n\n**What Data Will I Use in This Competition?**\n\nIn this competition, you will have access to two similar datasets that include attributes and labels. One dataset is titled `train.csv` and the other is titled `test.csv`. The datasets are anonymized.\nEach of these train and test have the same features, and only the train.csv file will include the correct labels, being the ransomware family.\n\nYou will have to model the train data, to the best of your ability and predict the corresponding labels in the test data. \n\nCheck out the “Data” tab to explore the datasets even further. Once you feel you have created a competitive model, submit it to see where your model stands on our leader board against other contestants.\n\n**Tips**\n\nWhile the score your model achieves is the most important factor in determining the Kaggle ranking, if two models achieves the same results, the earlier submission will rank higher!\nA tip from us is to do a quick baseline and submit periodically after improvements to the model. Use your 10 submissions well!\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Evaluation", "content": "**Goal**\nIt is your job to predict which ransomware family the Bitcoin address belongs to.\nFor each row in the test set, you must predict a class.\n\n**Evaluation**\n\nThe evaluation metric for this competition is Macro F1-Score. The F1 score, commonly used in information retrieval, measures accuracy using the statistics precision p and recall r. Precision is the ratio of true positives (tp) to all predicted positives (tp + fp). Recall is the ratio of true positives to all actual positives (tp + fn). The F1 score of a target class is given by:
\n\\[ F1 = 2\\frac{p \\cdot r}{p+r}\\ \\ \\mathrm{where}\\ \\ p = \\frac{tp}{tp+fp},\\ \\ r = \\frac{tp}{tp+fn} \\]
\nThe F1 metric weights recall and precision equally, and a good retrieval algorithm will maximize both precision and recall simultaneously. Thus, moderately good performance on both will be favored over extremely good performance on one and poor performance on the other. The Macro F1-score is calculated by averaging the F1-Score of every class
\nThe file should contain a header and have the following format:
\naddress,label", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "rules", "content": "This is a page where you can include rules that participants must accept before joining. You may wish to include rules like:\n\n - Don't cheat!\n - Apply yourself!\n - Have fun!", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false } ] }
17ujHj7k5sVXcL1JUd9WUv8nVfwuR1C4n1 ,0