introspector/zkprologml

zkPrologML数据集是一个完整的索引文件系统分析数据集，包含Monster Group分片分配信息。数据集包含8,017,192个索引文件，存储为250MB的parquet格式，包含16个列（如路径、Gödel数、分片、自然分类、特征向量和等）。数据集中的文件被分为71个Monster Group分片（0-70），自然分类分为五个类别（非常低、低、中、高、非常高），并提供了各类别的分布比例。Gödel数与分片之间存在完美的相关性（1.000），特征向量和与Gödel数之间的相关性为0.996。数据集还包含语义含义、使用描述和分类标签等信息。

The zkPrologML Dataset is a complete indexed filesystem analysis with Monster Group shard assignments. It contains 8,017,192 indexed files in a 250MB parquet format with 16 columns (including path, Gödel number, shard, natural class, eigenvector sum, etc.). The files are partitioned into 71 Monster Group shards (0-70) and classified into five natural classes (very_low, low, medium, high, very_high) with their distribution percentages provided. There is a perfect correlation (1.000) between Gödel numbers and shards, and a high correlation (0.996) between eigenvector sums and Gödel numbers. The dataset also includes semantic meaning, usage description, and classification labels.