Search for a command to run...
As the boost of the e-commerce industry, classifying users to achieve precise marketing has become an important topic. The traditional K-nearest neighbor algorithm is sensitive to outliers and has high computational complexity. Existing research lacks robust classification methods that can effectively handle the noise and high-dimensional features commonly present in e-commerce data while maintaining a low computational complexity. Therefore, the research proposes a new method integrating the Gaussian function into the KNN algorithm framework. The core innovation lies in the utilization of the probabilistic characteristics and smoothing properties of the Gaussian function, fundamentally changing the traditional KNN decision-making mechanism based on uniform distance. This method achieves a fundamental breakthrough in dealing with complex user behavior patterns by assigning probability-based weights to neighboring samples, thereby enhancing the model’s robustness to noisy data and local density variations. The experiment showcases that the model possesses a subject working feature area of 0.86 and 0.82 on the training and testing sets, respectively, which is significantly better than other models. This algorithm performs better in terms of convergence speed and average sum of squares of error compared to traditional algorithms, with an increase of 55.56% in convergence speed and a decrease of 71.74% in average sum of squares of error. In terms of customer feature classification, the algorithm significantly improves in terms of indicators. After strategic optimization, the marketing strategy forecast achieves high sales, and the overall revenue increases by approximately 20.74% compared to the traditional marketing strategy forecast. The research confirms the significant application value of the G-KNN algorithm in precise marketing of e-commerce: on the one hand, it can achieve more refined customer segmentation, providing a reliable basis for personalized recommendations and targeted promotions; On the other hand, by enhancing the accuracy of classification, marketing costs are effectively reduced and the input–output ratio of marketing are improved. This algorithm provides a practical technical tool for e-commerce platforms to optimize marketing decisions and enhance user satisfaction.