בדיקת נתיב סמנטי ספירלי: מיפוי תפוצה טכנולוגית לאורך קובצי עבודה אקדמיים, תיעודי פטנטים וגידול חומרה

תקציר מקורי באנגליתarXiv:2609.35845v1 Announce Type: new Abstract: Macroeconomic productivity metrics, such as Total Factor Productivity, register technological breakthroughs with multi-year reporting lags due to administrative survey intervals and national accounting conventions. This paper introduces Hyperspherical Semantic Trajectory Analysis (HSTA), an unsupervised quantitative methodology that tracks technology diffusion directly from unstructured scientific and commercial text streams. We analyze 30,000 filtered document records spanning academic preprints from arXiv and patent application records from the USPTO. By projecting high-dimensional Transformer sentence embeddings onto unit hyperspheres using Spherical K-Means clustering across eight primary sub-topics and UMAP manifold reductions, HSTA form