Affective Human-Machine Adaptation Method for Spinal Fixation Support Products

To address the intricate challenge of mapping product design elements in spinal fixation support to users' emotional adaptation satisfaction, we introduce a method for emotional human-machine adaptation driven by product reviews. We employ the Latent Dirichlet Allocation (LDA) model to cluster review data related to spinal fixation devices, thereby extracting essential user requirements. Simultaneously, we create a multidimensional design element space by incorporating visual data from the collected product samples. Subsequently, we encode the design elements of the collected samples and analyze the accompanying review data to gauge emotional satisfaction. The resulting codes and emotional scores are fed into a Backpropagation (BP) neural network, facilitating the construction of an emotional human-machine fit prediction model for spinal fixation braces. Ultimately, we assess the model's reliability using the K-fold cross-validation method. Additionally, we derive an optimized scheme for spinal fixation support.