Dominant Design Prediction with Phylogenetic Networks

Youwei He,Jeong-Dong Lee,Dawoon Jeong,Sungjun Choi,Jiyong Kim

2024-07-14

Abstract:This study proposes an effective method to predict technology development from an evolutionary perspective. Product evolution is the result of technological evolution and market selection. A phylogenetic network is the main method to study product evolution. The formation of the dominant design determines the trajectory of technology development. How to predict future dominant design has become a key issue in technology forecasting and new product development. We define the dominant product and use machine learning methods, combined with product evolutionary theory, to construct a Fully Connected Phylogenetic Network dataset to effectively predict the future dominant design.

Computational Engineering, Finance, and Science,Artificial Intelligence,Neural and Evolutionary Computing,Social and Information Networks

What problem does this paper attempt to address?

The key problem that this paper attempts to solve is: how to predict the future technology - dominant design. Specifically, the author proposes an effective method to predict the technology development trajectory from an evolutionary perspective. Product evolution is the result of technology evolution and market selection, and the formation of dominant design determines the trajectory of technology development. Therefore, how to accurately predict the future dominant design has become a key issue in technology prediction and new product development.

### Main research questions:
1. **How to construct phylogenetic networks in the technology evolution system?**
- The author proposes a Fully Connected Phylogenetic Network (FCPN) to overcome the limitations of traditional phylogenetic networks in technology evolution prediction.

2. **How to combine Graph Neural Networks (GNN) to predict future dominant products?**
- The author uses deep - learning models (especially the FCP - GNN framework) to predict which products may contain technologies (genotypes) that will become dominant in the future, thus identifying these dominant products in advance.

### Specific methods:
- **Data sources**: Data from the Chinese mobile phone market were used, covering 3,430 mobile phones from 2001 to 2021 and their 11,126 genotypes.
- **Data pre - processing**: Each product's attributes were converted into binary vectors representing their genotypes through one - hot encoding. Then the similarity between products in adjacent years was calculated to construct the FCPN.
- **Model construction**: A deep neural network model (DNN) was constructed based on the FCPN and GNN to predict which products may become dominant products in the future.

### Research objectives:
- **Predict dominant design**: By predicting which new technologies will become dominant in the future, help enterprises or countries make more informed strategic decisions.
- **Improve prediction accuracy**: By introducing the fully - connected phylogenetic network and graph neural network, improve the ability to predict future technology - dominant design.

### Key formula:
- **Jaccard similarity calculation**:
\[
\text{Similarity}(A, B)=\frac{|A\cap B|}{|A\cup B|}
\]
where \(A\) and \(B\) are genotype vectors of different products.

### Conclusion:
By constructing the FCPN and combining the GNN model, this study provides a new method to predict future dominant design, which has important guiding significance for the strategic decisions of enterprises or countries.