A novel algorithm (CAStream) is proposed to address the challenge of analyzing high-dimensional condensed matter structures by integrating mathematical probabilities with high-dimensional data. The algorithm improves the clustering accuracy and memory efficiency of high-dimensional data streams through subspace clustering. Taking nylon 6 (PA6) as the research object, combined with X-ray diffraction and differential scanning calorimetry, the effectiveness of the algorithm in the evolution of crystal structure (α/γ crystal transformation) was verified. The research results show that the algorithm has been tested for its ability and scalability in processing data of different spatial dimensions, with a spacing of 5 between datasets. A simulated dataset with different clustering dimensions was established to test the effectiveness of the algorithm in different clustering dimensions of the data stream. It was found that the algorithm performed well in the spatial aspect of the data stream, had good scalability, and performed well for any clustering. It can effectively handle high-dimensional data and has higher accuracy than the CluStream algorithm due to its use of subspace clustering. Nylon 6 mainly exhibits different characteristic peak shapes, such as the alpha crystalline characteristic peak shape at 2θ = 19.8° and θ = 22.9°, and the gamma crystalline characteristic peak shape corresponding to the point at θ = 20.8°. In addition, diffraction peaks of the alpha crystal form were observed, attributed to the addition of apigenin in the range of 0.01–0.05%. The effect of apigenin on the transformation of PA6 from the γ crystal form to the α crystal form decreases as its amount increases. The diffraction peak of the γ crystal structure in the figure appears again at θ = 21.2, during which 1% apigenin is added. It is evident that the suggested approach can successfully and efficiently generate the condensed structure. The research provides a new paradigm for the regulation of polymer condensed matter structure and the design of high-performance materials.