Tribler Learning-to-Rank Dataset

The Tribler Learning-to-Rank (LTR) Dataset is a collection of feature vectors derived from a decentralized search engine conducted within the Tribler peer-to-peer file-sharing network. This dataset was created to train and evaluate DART, an LTR-based algorithm for relevance ranking in decentralized file-sharing systems.The dataset supports research into decentralized search and ranking techniques by providing feature vectors extracted from real user interactions. It allows researchers to develop, validate, and compare LTR models for the decentralized setting.<br>This dataset comprises rows containing a binary relevance judgment, a query ID, and 31 normalized features.<br>Example:```1 qid:8030 0:0.311253027575435 1:0.3623667000954467 2:0.2640038215062309 3:0.2908247887862055 4:0.0583718121596237 5:-0.18465409343274 6:1.327852667807301 7:1.315861763360352 8:1.509079253500692 9:-0.1600325676141462 10:-0.5992397181984241 11:0.5291165364861627 12:0.4186126253353217 13:0.4601304198871708 14:0.2818017229171959 15:-0.237928002400229 16:0.1731179222782891 17:0.0738413823487902 18:0.3099197449886553 19:0.1805433703801239 20:0.1158592011796661 21:-0.665142871178777 22:0.3328536662895701 23:0.2948988525040041 24:0.7315729714005331 25:-0.6902133277846004 26:2.62527037747493 27:-0.3083199080468793 28:0.5082540870914015 29:-1.084397975003081 30:-5.408917243099888<br>0 qid:8030 0:0.2625962288699837 1:0.3623667000954467 2:0.2640038215062309 3:0.2908247887862055 4:0.0583718121596237 5:-0.18465409343274 6:1.327852667807301 7:1.315861763360352 8:1.509079253500692 9:-0.1600325676141462 10:-0.5992397181984241 11:0.5291165364861627 12:0.4186126253353217 13:0.4601304198871708 14:0.2818017229171959 15:-0.237928002400229 16:0.1994409565033298 17:0.0738413823487902 18:0.3099197449886553 19:0.02724546833125713 20:0.4530615272328508 21:-0.665142871178777 22:0.3328536662895701 23:0.2948988525040041 24:0.6189088582085388 25:-0.6902133277846004 26:-0.3433711973076669 27:-0.3083199080468793 28:0.5082540870914015 29:-1.637896085379517 30:-2.427152848520973```<br>Please cite the following article when using this dataset:**to be added**Tribler 排序学习（Learning-to-Rank, LTR）数据集是一组从运行于Tribler点对点（peer-to-peer）文件共享网络中的去中心化搜索引擎中提取得到的特征向量集合。本数据集专为训练与评估DART而构建，DART是一种应用于去中心化文件共享系统相关性排序任务的基于LTR的算法。本数据集通过提供从真实用户交互中提取的特征向量，为去中心化搜索与排序技术的相关研究提供支持，可帮助研究人员开发、验证并对比适用于去中心化场景的LTR模型。<br>本数据集由多行数据构成，每行包含一个二元相关性标签、一个查询ID（query ID）以及31个归一化特征。<br>示例：1 qid:8030 0:0.311253027575435 1:0.3623667000954467 2:0.2640038215062309 3:0.2908247887862055 4:0.0583718121596237 5:-0.18465409343274 6:1.327852667807301 7:1.315861763360352 8:1.509079253500692 9:-0.1600325676141462 10:-0.5992397181984241 11:0.5291165364861627 12:0.4186126253353217 13:0.4601304198871708 14:0.2818017229171959 15:-0.237928002400229 16:0.1731179222782891 17:0.0738413823487902 18:0.3099197449886553 19:0.1805433703801239 20:0.1158592011796661 21:-0.665142871178777 22:0.3328536662895701 23:0.2948988525040041 24:0.7315729714005331 25:-0.6902133277846004 26:2.62527037747493 27:-0.3083199080468793 28:0.5082540870914015 29:-1.084397975003081 30:-5.408917243099888<br>0 qid:8030 0:0.2625962288699837 1:0.3623667000954467 2:0.2640038215062309 3:0.2908247887862055 4:0.0583718121596237 5:-0.18465409343274 6:1.327852667807301 7:1.315861763360352 8:1.509079253500692 9:-0.1600325676141462 10:-0.5992397181984241 11:0.5291165364861627 12:0.4186126253353217 13:0.4601304198871708 14:0.2818017229171959 15:-0.237928002400229 16:0.1994409565033298 17:0.0738413823487902 18:0.3099197449886553 19:0.02724546833125713 20:0.4530615272328508 21:-0.665142871178777 22:0.3328536662895701 23:0.2948988525040041 24:0.6189088582085388 25:-0.6902133277846004 26:-0.3433711973076669 27:-0.3083199080468793 28:0.5082540870914015 29:-1.637896085379517 30:-2.427152848520973<br>使用本数据集时，请引用以下文献：**待补充**