Quantum Recurrent-Assisted Adaptive K-harmonic Means Clustering Approach for Classification Task

The rapid growth of data and the collective complexity of high-dimensional classification issues have pushed traditional computer technologies to their limits. However, quantum computing has emerged as a possible paradigm to address these difficulties. In this study, a novel quantum machine learning procedure for classification tasks is proposed. The research is separated into three phases: Pre-processing, clustering and classification. Initially, collect the data from the dataset and perform pre-processing using min-max normalization. Clustering and classification methods are essential constituents of this research because they enable the efficient organization and interpretation of complex datasets related to power generation and other features. Novel clustering methods are used to group similar data points based on their characteristics without requiring labelled data. This clustering approach has been performed by quantum assisted adaptive k-harmonic means (Quant-AdKharmonic) algorithm, which can group similar data effectively. After a clustering approach-based label creation, the classifier model, namely Entanglement-Variational Quantum Recurrent Classifier, is used for the classification task (En-VarQuantum). The integration of both models has handled noisy data for better power prediction and analysis. The performance of the suggested model is assessed employing two datasets: TWTDUS and SDWTT18. Thus, the findings of the TWTDUS dataset, utilizing clustering and classification methods, achieve the highest accuracy of 97.18%. Similarly, for the SDWTT18 dataset, the proposed method achieves an accuracy of 98.17%.