Integration of quantum computing models into data analysis pipelines

With the increasing complexity and dimensionality of data, traditional machine learning methods face limitations in expressive power, scalability, and computational efficiency. In this paper, we consider an approach to building hybrid data analysis pipelines that integrate classical and quantum computing models. Using the PennyLane quantum computing framework and the PyTorch and Sciki-learn machine learning libraries, architectures for classification tasks have been developed, the key element of which are variational quantum circuits integrated into a classical pipeline with feature preprocessing and measurement postprocessing modules. In the course of the experimental study, a comparative analysis of classical and hybrid models was carried out on samples of different volumes using three alternative entanglement topologies in quantum circuits. The results obtained confirm the competitive accuracy of hybrid models and indicate their promise for working with balanced datasets and in conditions of limited sample size.