Thermal Error Transfer Prediction Modeling of Machine Tool Spindle with Self-Attention Mechanism-Based Feature Fusion

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The overall concept of the paper is innovative, particularly the shift from using multiple temperature points to focusing on key temperature points. While the superiority of the model is difficult to definitively prove (when compared with other relevant studies), the explanation provided is supported by substantial scientific evidence. The study compares training times of various learning methods using tables, though the comparison is mainly conducted with larger models. Recent studies indicate that traditional machine learning models sometimes offer advantages in terms of training time and model configuration. Nevertheless, this study contributes significantly to the field.

1. The representations in Figures 2 and 3 are not sufficiently clear, especially in terms of making it intuitive to read the data on the Y-axis.

2. The study does not provide a detailed explanation for the use of five displacement sensors. This issue may lead to differences in the outcomes of subsequent experimental designs.

3. The study lacks information regarding the models and specifications of the temperature and displacement sensors.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors proposes a thermal error model that utilizes a fusion of temperature features with self-attentive weights to improve accuracy and efficiency. First, sensitive temperature data is fused by using a self-attention mechanism to enhance information extraction for accurate thermal error modeling. Second, an improved adaptive matrix method based on direct normalization bridges the gap between source and target domains, using an auto-encoder to effectively transfer knowledge. Last, the non-parametric method aligns feature distributions between domains, showing better performance in thermal error predictions compared to other existing methods. Overall the proposed method enhances training speed and prediction accuracy, resulting in more stable thermal error predictions. I think this paper is well-written, and the proposed method is sound. However, I do have the following concerns before the consideration of publication: 

1. 29 out of 30 references are authored by Chinese scholars. I do suggest adding more citations from other international countries. 

2. Based on the iThenticate report, there is an overlap starting from line 198 to line 209 with the referred paper. Authors are suggested to paraphrase it to avoid plagiarism. 

Li, Dengshan, and Lina Li. "Novel Hybrid Calibration Transfer Method Based on Nonlinear Dimensionality Reduction for Robust Standardization in Near-Infrared Spectroscopy." Analytical Letters 56, no. 15 (2023): 2522-2539.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf