Pioneering Technology Mining Research for New Technology Strategic Planning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The importance and originality of the study and the summary of the chapters should be presented.
The method section is adequate and detailed. 
The findings section is explanatory and detailed. 
The preparation of the figures is careful. 
While mentioning the aims of the study, it should be mentioned which deficiencies this study will fill and its originality.
The number of references in the literature section is very few and insufficient. 
Table 5 and Table 6 should be explained and interpreted in detail.
The discussion section is insufficient. In this section, the evaluation of the study should be made by comparing it with the literature. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

It should be reviewed.

Author Response

Comments 1:

The importance and originality of the study and the summary of the chapters should be presented.

Response 1:

1. Thank you very much for your suggestion. We have added an explanation of the importance and originality of this study in the appropriate section of the introduction. Elaborated on how this research fills the gap in existing technology strategic planning research, as well as its novelty in theory and practice. Specifically, as follows:

To address these challenges, this study introduces a Multi-Dimensional Robust Stacking (MDRS) Model, which innovatively handles patent data, enabling precise identification of original innovative technologies and overcoming the limitations of existing methods in dealing with anomalous data and complexity. The MDRS model enhances the accuracy and robustness of technology predictions through comprehensive patent data mining and analysis. Its multi-stage analysis method effectively handles data anomalies and provides a thorough quantitative analysis of technology trends, offering theoretical and strategic insights for governmental and enterprise technological development plans.

Modification location: Section 1 Paragraph 6.

2. Thank you for your suggestion. We have added the themes of each chapter and elaborated on the main structure of the article in the last paragraph of the introduction. Specifically, as follows:

The remainder of this paper is organized as follows: the second part reviews the research on pioneering technologies, the role of patent analysis in technology trend forecasting, and the shortcomings of existing studies. The third part then elaborates on the purpose, methods, and four stages of constructing the MDRS model. The fourth part provides a detailed discussion of data sources, selection criteria, and experimental results, demonstrating the application effects of the MDRS model in predicting 3D printing technology and comparing it with traditional models. Finally, the study discusses the theoretical contributions of the MDRS model in the field of technology forecasting and innovation, emphasizing its practical application value in guiding technology strategic planning and promoting the development of 3D printing technology. The study also points out the limitations and future research directions.

Modification location: Section 1 Paragraph 10.

Comments 2:

While mentioning the aims of the study, it should be mentioned which deficiencies this study will fill and its originality.

Response 2:

Thank you for your valuable suggestion. In the appropriate section of the introduction, we have discussed the shortcomings filled in this study and its originality. Specifically, as follows:

Existing research primarily focuses on the qualitative analysis of technology, lacking comprehensive quantitative descriptions of cutting-edge technologies and typically identifying them based on application prospects alone. This study fills this gap by constructing a technological development matrix based on the MDRS model results, analyzing the leading positions of upstream, midstream, and downstream technologies within the industry chain [25]. This method accurately identifies foundational innovative technologies across various sectors, providing deeper strategic insights and aiding governments and enterprises in formulating appropriate technological development plans.

Modification location: Section 1 Paragraph 9.

Comments 3:

The number of references in the literature section is very few and insufficient.

Response 3:

Thank you for your valuable comments. Regarding the issue of "insufficient citation numbers in the literature section," we have made modifications and supplements. The current literature mostly analyzes pioneering technologies from a macro perspective. We have supplemented the paper with relevant content to discuss the importance of technology's inherent characteristics, and increased the depth of discussion on technology characteristics. Furthermore, to compensate for the limitations of existing methods in early technology identification, we have introduced the latest research on patent clustering and machine learning techniques, further emphasizing the need for more methods and tools to fill the gaps in existing research. We have also increased the number of additional literature citations to make the reference list more comprehensive and broadly representative, providing a more comprehensive academic background support for the study. These supplements have enhanced the scientific and cutting-edge nature of the paper, making its content more authoritative and persuasive. Thank you again for your suggestions.

The specific modifications are as follows:

Literature review

Pioneering technologies refer to groundbreaking technologies that hold a central position in their field, significantly influencing other technologies and contributing substantially to technological development. In other words, pioneering technologies are indispensable within their specific domains [26]. They play a crucial role in industrial diversification and stable development. Currently, the analysis of pioneering technologies primarily focuses on their qualitative characteristics. For example, at the content level, the construction of characteristic frameworks mainly relies on macro-background factors, such as market conditions, rather than on micro-characteristics like technology [27]. Recently, quantitative analysis of patents has garnered scholars' attention for the study of pioneering technologies. The citation relationships among patents can be objectively visualized through patent networks, allowing the measurement of each node's core degree within the network [28]. One study explored the dynamic process of technological innovation by analyzing the relationship between collaboration networks and structural holes [29]. Additionally, patent citation networks have been used to identify emerging technologies by analyzing citation relationships and determining the importance of specific patents within the network [30]. However, existing literature mainly analyzes pioneering technologies from a macro perspective, from the market. Most methods are suitable only for identifying pioneering technologies in the growth and maturity stages, not in the nascent stage. Additionally, most research focuses on the impact of technology on the market from a corporate perspective, neglecting the characteristics of the technology itself [31].

Patent research enables a comprehensive analysis of trends in technological innovation, aiding companies and governments in developing strategic plans for research and proposing new commercial technology products, thereby enhancing market competitiveness [32]. For instance, patent data can predict the development trends of photovoltaic technology, thus providing a scientific basis for companies to formulate their R&D strategies [33]. Most current patent mining methods use a fixed mining model to identify the dominant role of patents, ignoring the impact of anomalous pioneering technologies and long analysis cycles on predictive model performance [34]. Although existing methods can analyze the paths of technological innovation and evolution through patent citation networks [35], they exhibit significant limitations when it comes to identifying nascent technologies. For example, the study found that, although patent citation networks can effectively analyze technological innovation paths, there are still deficiencies in early-stage technology identification, necessitating more methods and tools to address this gap [36]. In addition, machine learning and genetic algorithms were used to validate the effectiveness of the model, providing reference for improving the accuracy of technical recognition [37]. Therefore, identifying pioneering technologies in their early development stages is not only feasible but also necessary.

To overcome these shortcomings, this study proposes the MDRS model. This model extracts highly robust leading indicators that comprehensively reflect the characteristics of pioneering technologies, allowing for accurate identification of pioneering technologies in the early stages of technological development. By using this model, this study can not only identify and predict emerging technology directions but also provide scientific strategic planning bases for technology development for companies and governments, effectively enhancing market competitiveness and technological innovation capabilities.

Modification location: Section 2 Paragraph 1-3.

Comments 4:

Table 5 and Table 6 should be explained and interpreted in detail.

Response 4:

We have revised the discussion related to Table 5 to provide a more detailed explanation：“As shown in Table 5, the average MSE of the MDRS model is the lowest compared to conventional models, indicating that the prediction based on hyper-robust leading indicators is more accurate. Specifically, the LTDMM model achieved the lowest MSE of 0.28, whereas the SVM model had the highest MSE of 1.33. Other models, including various regression models and BPNN, had intermediate MSE values ranging from 0.47 to 0.80.”

Similarly, we have expanded the discussion of Table 6 to offer a more comprehensive interpretation：“Table 6 reveals that the maximum, average, and minimum MSE of the constraint nested MDRS model are lower than those of the NO-MDRS model. Specifically, the maximum MSE for the constraint-nested MDRS model is 0.43, while it is 0.56 for the NO-MDRS model. The average MSE for the constraint-nested model is 0.28 compared to 0.47 for the NO-MDRS model. Lastly, the minimum MSE for the constraint-nested model is 0.24, whereas it is 0.35 for the NO-MDRS model.”

We trust that these expanded discussions provide the necessary detail and clarity to fully address the reviewer’s request for a more thorough explanation of the data presented in Tables 5 and 6.

Modification location: Section 4.3, Paragraph 3,5.

 Comments 5:

The discussion section is insufficient. In this section, the evaluation of the study should be made by comparing it with the literature.

Response 5:

Thank you for the valuable feedback. Based on your suggestion, we have revised the discussion section to include a more comprehensive evaluation by comparing our study with existing literature. Our proposed MDRS model offers several advantages over existing methods: more precise prediction of frontier technology trends, by constructing single, robust, and hyper-robust indicators, we significantly improve the accuracy and robustness of technology trend predictions, especially in patent data analysis; emphasis on original innovation and frontier technology analysis, unlike other studies that focus on predicting opportunities for technological convergence, our research emphasizes the mining of frontier technologies, providing more targeted technological development strategies for enterprises and governments; multi-dimensional indicator analysis, compared to methods that rely on single-type information, our approach integrates multiple patent indicators, enhancing the comprehensiveness and accuracy of technology path mapping. In summary, the MDRS model significantly enhances the accuracy and robustness of technology trend prediction, providing a scientific basis for optimizing resource allocation and enhancing market competitiveness. By comparing our model with existing research, we demonstrate its advantages and improvements in patent data analysis.

The specific modifications are as follows:

Conclusion and implications

Breakthrough innovations stem from original innovation research, which is a crucial method for helping enterprises gain a competitive edge in the fiercely competitive market. The core of original innovations lies in identifying frontier technologies that lead industry innovation and technological expansion. By early identification and screening of leading technologies and formulating correct development plans for original innovative technologies, companies, and governments can effectively reduce misjudgments regarding technological directions in the innovation process, thereby improving the efficiency of technological innovation. However, due to the scarcity of leading technologies, their metric values significantly differ from other technologies, displaying outlier characteristics, which poses substantial challenges to the prediction of pioneering technologies.

Therefore, this study proposes a pioneering technology mining method, that delves deeply into pioneering technologies to reduce misjudgments by enterprises and governments during the innovation process and improve innovation efficiency. Compared with the study by Kim G and Bae J [72], which suggests a novel approach to forecasting promising technology using patent analysis, our research presents a Multi-Dimensional Robust Stacking (MDRS) model that can more precisely predict frontier technologies. Additionally, our research emphasizes original innovation and frontier technology more than the approach by Park I and Yoon B [73], who explore technological opportunities for convergence by predicting potential technological knowledge flows between heterogeneous fields, our MDRS model delves deeper into analyzing and mining frontier technologies from patent data, providing more targeted technological development strategies for enterprises and governments. Furthermore, compared to Liu H et al. [74], who developed a method for mapping the technology evolution path using a non-parametric topic model, our study incorporates multiple patent indicators to improve the comprehensiveness and accuracy of technology path mapping. By proposing the MDRS model, this research systematically enhances the accuracy and robustness of technology trend predictions, particularly in the analysis of patent data. It achieves the construction of single, robust, and hyper-robust indicators. These indicators, through comprehensive patent data analysis, describe the leading characteristics of technologies and overcome the limitations of traditional analysis methods.

Theoretical implications

The innovative MDRS model presented in this study represents a significant advancement over traditional patent data analysis methods. Conventional approaches often rely on single-dimensional indicators and are susceptible to outlier data, resulting in lower prediction accuracy. For instance, some studies have highlighted the limitations of traditional single-dimensional models in capturing the complexity and dynamism of technology evolution, which MDRS effectively addresses [75][76]. MDRS model overcomes these limitations by integrating multi-dimensional single-leading indicators to construct robust composite indicators, further enhanced through non-linear fusion using BPNN. This approach markedly improves both the accuracy and robustness of predictions, providing a novel methodology for quantitative analysis of technology trends. Furthermore, by analyzing technology leadership across dimensions such as technology popularity, authority, and the inventor's social relationships, this study uncovers the intrinsic mechanisms of technological leadership at various stages and application scenarios. This multi-dimensional analysis enriches the theoretical framework of technological innovation and leadership, offering a solid foundation for future research. The work of Fleming et al. on the importance of inventor networks in technological advancement further supports our findings on social relationships as a critical dimension [77]. Additionally, the study underscores the importance of interdisciplinary cooperation in technological strategic planning, introducing new methods and tools for integrating and applying data across multiple disciplines, thereby fostering effective resource allocation and collaborative innovation.

Modification location: Section 5, Paragraph 1-2 and section 5.1.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Comment on “Pioneering Technology Mining Research for New Technology Strategic Planning ”

 

The main work conducted in this thesis involves the development and validation of a novel method known as the Multi-Dimensional Robust Stacking (MDRS) model, designed to mine and identify pioneering technologies from patent data. The specific tasks undertaken include:

1. Model Development: The thesis introduces the MDRS model, which analyzes patent data through four distinct phases: single indicator construction, robust indicator mining, hyper-robust indicator construction, and pioneering technology analysis.

2. Indicator Creation: A series of single leading indicators were developed based on dimensions such as technology popularity, authority, and inventor's social relationships (ISR). These indicators were then fuzzified to enhance the model's robustness against outliers.

3. Data Mining and Integration: Various data mining techniques, including regression analysis, Support Vector Machine (SVM), and neural network models, were employed to integrate and combine single indicators into more robust composite indicators.

4. Model Training and Validation: A Backpropagation Neural Network (BPNN) was used to nonlinearly integrate the indicators, creating hyper-robust indicators. The model was trained and validated using patent data from the 3D printing industry.

5. Technological Development Matrix: Based on the outcomes of the MDRS model, a technological development matrix was constructed to analyze and evaluate the position and impact of different technologies within the industry chain.

6. Strategic Planning Support: The thesis provided scientific support and operational guidelines for technology strategic planning in the 3D printing industry through the technological development matrix.

7. Empirical Study: The MDRS model was applied to the 3D printing industry to demonstrate its effectiveness in practical technology forecasting and strategic planning.

8. Theoretical and Practical Contributions: The thesis enriched the theoretical framework of technological innovation and leadership and offered practical support for the growth of emerging industries and the enhancement of market competitiveness.

9. Future Research Directions: The thesis concluded by identifying limitations and suggesting directions for future research, such as in-depth analysis of patent text content and integration of multiple data sources to improve the accuracy and generalizability of technology forecasting.

 

Here are some of my suggestions

1. Were the authors able to test the generalization capabilities of the MDRS model across different technology domains and industries to ensure broad applicability of the model?

2. It is recommended that the authors provide more specific policy recommendations and implementation strategies, including evaluation mechanisms, to help governments and enterprises better formulate and implement technology development plans.

3. Most of the references are quite old, please add some recent literature on related topics.

4. This article has some spelling mistakes and illogical statements, it is recommended to find a professional touch-up agency to improve the language quality of this article.

 

Comments for author File: Comments.pdf

Author Response

Comments 1:

Were the authors able to test the generalization capabilities of the MDRS model across different technology domains and industries to ensure broad applicability of the model?

Response 1:

First and foremost, we sincerely appreciate your valuable time and meticulous review. Your question regarding the generalization capability test of the MDRS (Multi-Dimensional Robust Stacking) model across different technological fields and industries is a critical factor for ensuring the model's broad applicability.

In this study, we focused on patent data in the 3D printing technology sector to test the effectiveness of the MDRS model. The 3D printing industry was chosen as a case study because it is a rapidly evolving field filled with innovative opportunities, providing us with an ideal testing platform. By deeply analyzing the core technologies at different stages of the 3D printing industry's value chain, we constructed a technology development matrix and conducted an empirical analysis of the MDRS model.

Although this study primarily focuses on the 3D printing industry, the design principles and methodologies of the MDRS model have cross-sector applicability. The model's multi-dimensional analysis framework and robust metric construction methods theoretically can adapt to different technological fields and industries. However, we acknowledge that to fully verify the model's generalization capability, future research should test it across more technological fields and diverse industries.

We agree that cross-sector testing will provide further evidence of the model's broad applicability. To this end, we plan to extend the application range of the MDRS model in future research, including but not limited to other industries such as biotechnology, information technology, energy, and materials science. This will help us better understand the model's performance in different contexts and conduct a more comprehensive evaluation of its generalization capability.

Comments 2:

It is recommended that the authors provide more specific policy recommendations and implementation strategies, including evaluation mechanisms, to help governments and enterprises better formulate and implement technology development plans.

Response 2:

In response to the reviewer’s this comment, this study have revised section 5.2 Practical Implications to include more specific policy recommendations and implementation strategies, along with evaluation mechanisms.

This revision focuses on two primary aspects of our research’s practical significance: (1) identifying core technologies at various industry stages to ensure rational resource allocation, and (2) focusing on the 3D printing industry to provide insights into technology trends and strategic value, supporting the growth of emerging industries.

Our updated content now offers detailed implementation strategies and corresponding evaluation mechanisms as the newest manuscript.

5.2 Practical implications

The application of the MDRS model provides precise guidance for technology strategic planning, demonstrated by the 3D printing technology strategic planning support matrix. This matrix offers a clear theoretical basis for formulating development strategies, helping to identify core technologies at various industrial stages. This guidance enables rational resource allocation, facilitating feasible technology plans and enhancing market competitiveness. Echoing the practical implications discussed by McAfee et al., in their study on strategic planning in the AI industry, our model’s application also underscores the necessity of tailored strategies for different technological contexts [78].

Moreover, the model’s ability to identify long-term leading original innovation technologies reduces misjudgments in technology direction, improving the efficiency of innovation processes for enterprises and governments. This aligns with the conclusions of Rohrbeck et.al. [79], who emphasized the value of accurate technology forecasting in minimizing strategic risks. By focusing on the 3D printing industry, the study offers insights into technology trends and strategic value, supporting emerging industries' growth. For the 3D printing industry, the government should formulate a special development plan, promoting the coordinated development of upstream and downstream enterprises in the 3D printing industry chain, forming a complete industrial ecosystem. Implementation strategies include using the MDRS model to regularly analyze 3D printing technology trends, creating differentiated policy support systems for various stages of development (e.g., R&D support, market promotion, talent introduction), thereby promoting industrial clustering and collaborative innovation. The evaluation mechanism should establish a set of industry development indicators for the 3D printing sector (such as technological innovation capability, market size, and completeness of the industry chain), regularly assessing the industry's development status and adjusting and optimizing policy measures based on evaluation results.

All in all, the findings assist governments in formulating policies that support original innovation technologies, optimizing tax incentives, R&D funding, and other measures, ultimately fostering technological innovation and economic growth in enterprises.

Modification location: Section 5.2.

Comments 3:

Most of the references are quite old, please add some recent literature on related topics.

Response 3:

Thank you for your valuable suggestion. We have added the latest literature related to the topic in appropriate places in the article.

Previous references:

[6] Jun, S.; Sung Park, S. Examining technological innovation of Apple using patent analysis. Industrial Management & Data Systems 2013, 113, 890-907.

[8] Magruk, A. Innovative classification of technology foresight methods. Technological and Economic Development of Economy 2011, 4, 700-715.

[18] Breiman, L. Bagging predictors. Machine Learning 1996, 24, 123-140.

[19] Freund, Y., & Schapire, R. E. A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences 1997, 55(1), 119-139.

[29] Campbell, R.S. Patent trends as a technological forecasting tool. World Patent Information 1983, 5, 137-143.

[35] Shiu-Wan, H.; An-Pang, W. A small world in the patent citation network. In Proceedings of the 2008 IEEE International Conference on Industrial Engineering and Engineering Management, 8-11 Dec. 2008, 2008; pp. 1-5.

[53] Lecun, Y.; Bottou, L.; Bengio, Y.; Haffner, P. Gradient-based learning applied to document recognition. Proceedings of the IEEE 1998, 86, 2278-2324.

Revised references:

[6] Barley B, Kitamura A, Loar T, et al. An Investigation of the Motivations and Strategies Behind Apple’s Product Design. Innovation Management in the Intelligent World: Cases and Tools, 2020, 3-27.

[8] Akbari M, Khodayari M, Khaleghi A, et al. Technological innovation research in the last six decades: a bibliometric analysis. European Journal of Innovation Management, 2021, 24(5), 1806-1831.

[18] Jun S P, Lee J S, Lee J. Method of improving the performance of public-private innovation networks by linking heterogeneous DBs: Prediction using ensemble and PPDM models. Technological Forecasting and Social Change, 2020, 161, 120258.

[19] Natekin A, Knoll A. Gradient boosting machines, a tutorial. Frontiers in neurorobotics, 2013, 7, 21.

[32] Jiaxuan W. Link prediction of directed knowledge network based on patent co-citation. 2023 IEEE 3rd International Conference on Data Science and Computer Application (ICDSCA). IEEE, 2023, 21-27.

[37] Li S, Zhu B, Zhang Y, et al. A Two-Stage Nonlinear User Satisfaction Decision Model Based on Online Review Mining: Considering Non-Compensatory and Compensatory Stages. Journal of Theoretical and Applied Electronic Commerce Research 2024, 19(1), 272-296.

[39] Balland P A, Rigby D, Boschma R. The technological resilience of US cities. Cambridge Journal of Regions, Economy and Society, 2015, 8(2), 167-184.

[57] Ciresan D C, Meier U, Masci J, et al. Flexible, high performance convolutional neural networks for image classification. Twenty-second international joint conference on artificial intelligence. 2011.

Modification location: Section 3.1 and 3.4.

Comments 4:

This article has some spelling mistakes and illogical statements, it is recommended to find a professional touch-up agency to improve the language quality of this article.

Response 4:

Thank you for your valuable suggestion. We have carefully checked the spelling and logical statements in the article, and polished the language quality.