*2.2. Methodology*

This study utilizes association strength as the technological relatedness measure between TIs of patent [30], and betweenness centrality as a priority measure of TI [31]. The calculation process of betweenness centrality was performed by using the software package "networkX" [32].

#### 2.2.1. Technological Relatedness: Association Strength

In this study, we calculated the association strength similarity. The calculation processes of the two measures were undertaken by following the formulas provided below [30].

Equations (1)–(5) show the calculation process of association strength similarity.

$$
\varepsilon t i\_{ij} = \sum\_{p=1}^{m} t i\_{p\bar{i}} t i\_{p\bar{j}\prime} \quad \text{for } \bar{i} \neq j,\tag{1}
$$

In Equation (1), *ctiij* is the number of co-occurrence of TIs *i* and *j*. *tipi* and *tipj* are terms for counting the number of TIs in the patent. *tip*· is one (*tip*· = 1) when the TIs *i* or j exist in patent *p*, and *tip*· is zero *tip*· = 0 when either TIs do not exist in patent *p*. Thus, *ctiij* becomes one when both *tipi* and *tipj* are one. *m* is number of total patents for a five-year research period.

$$\text{sti}\_{i} = \sum\_{j=1, j \neq i}^{m} \text{cti}\_{ij} \,\,,\tag{2}$$

$$\text{sti}\_{\dot{j}} = \sum\_{i=1, j \neq i}^{n} \text{cti}\_{i\dot{j}\\_\ast} \tag{3}$$

$$T = \sum\_{i=1}^{n} \sum\_{j=1}^{n} c t i\_{ij} \, \prime \, \tag{4}$$

In Equations (2) and (3), *stii or j* is the total co-occurred number of TIs *i* or *j* for a five-year research period. *T* is sum of total co-occurred number of whole TIs.

$$S\_{C\_{ij}} = \frac{\frac{cti\_{ij}}{T}}{\frac{sti\_i}{T}\frac{sti\_j}{T}}\,'\,\tag{5}$$

In Equation (5), *SCij* is the similarity measure between TIs *i* and *j*, which occur over a 5-year research period.
