A Study on Clustering Analysis Method Based on Multi-dimensional Feature Engineering Fusion

To address the limitations of current clustering analysis methods, particularly the reliance on a single feature engineering approach and the limited precision in identifying data objects, this paper proposes a clustering method based on multi-dimensional feature engineering. The proposed approach integrates both empirical and theoretical techniques into a unified feature extraction framework to enable comprehensive multi-dimensional data representation. It combines deep learning models and the isolation forest algorithm as preprocessing steps, followed by K-means clustering for accurate object categorization. Experimental results demonstrate that the method achieves high-quality classification outcomes, offering a systematic solution for data mining and contributing both theoretical insights and practical value to the fields of machine learning and pattern recognition.