CCVRL Structured Datasets for Context-Conditional Value Relation Learning in Ideological and Political Education

<p dir="ltr">This dataset package accompanies the study “Context-Conditional Value Relation Learning for Case Analysis in Ideological and Political Education.” It provides structured machine-learning data for contextual value recognition from Chinese occupational-behavior texts and normative value rules.</p><p dir="ltr">The package contains two CSV datasets: CoreValue.csv and C_VARC_8.csv. Both datasets use a unified eight-category value inventory consisting of civility, justice, equality, rule of law, patriotism, dedication, integrity, and friendliness.</p><p dir="ltr">CoreValue contains 6,994 occupational-behavior samples.

The dataset follows the experimental split used in the study, with 5,595 training samples, 699 validation samples, and 700 test samples. It supports multi-label value-category recognition and, where annotations are available, prediction of activation or inhibition directions. The structured fields also represent the event-role information used by the CCVRL framework, including subject, action, precondition, object, manner, time, and location.</p><p dir="ltr">C-VARC-8 represents the value-recognition task derived from the C-VARC Chinese value-rule resource.