Non-face-to-face life is becoming common through COVID-19 pandemic, and demand for purchasing products through e-commerce is increasing. In line with these changes in the era, the last mile sector is drawing attention, and nationally, private companie...
Non-face-to-face life is becoming common through COVID-19 pandemic, and demand for purchasing products through e-commerce is increasing. In line with these changes in the era, the last mile sector is drawing attention, and nationally, private companies are also trying to increase their competitiveness through the development of related fields. The key to strengthening competitiveness is to provide information to predict promising future technologies and establish a reasonable technology strategy with limited resources, which will lead to a technological advantage.
Technology prediction was generally conducted through the gathering of opinions from a group of experts, but it has a limitation that it takes time and money and is subjective. Recently, studies have been conducted to minimize the threshold using a combination of quantitative analyses. Among them, analysis using patent information representing technical elements is the most reasonable, and various research and analysis methods are used to conduct technology prediction. In this study, we reviewed existing studies to confirm that the lack of technical prediction research in the last mile field and that technology prediction methods using patent information are most reasonably utilized. Therefore, I conducted the study with the goal of predicting promising technologies in the field of Last Mile, which applied the most reasonable patent analysis technology prediction method.
Patent information was collected through the Wipson database for patents released 18 months after applications in Korea, China, the United States, and Europe from January 1974 to August 2020. The invention name and summary information of the collected patents were formatted through the text mining process, and the top 500 keywords were extracted based on the frequency of their appearance. The extracted keywords were able to be selected as many as five groups that did not overlap keywords as possible through the technology clustering process through the LDA. Ten key keywords for each selected group could be selected to define each technical field.
ARIMA time series analysis was conducted to confirm the technology trends of each group. As a result of checking the trend of future patents over time and confirming future prospects, all five groups showed an upward curve, indicating a high possibility of future development. Gap nodes were derived through GTM patent map analysis for groups that were confirmed to be highly promising, and blank nodes were identified through keyword-patent matrix, and promising technologies by group can be predicted through key keywords belonging to each node.
Based on research and analysis results, promising technologies in each technology field were "Automated operation technology using robots", "Aids related to goods classification and delivery", "Low-cost classification and transport optimization technology", "Monitoring technology for safe delivery and management", and "Structures and facilities for classification in logistics centers".
As technology's influence has increased, preparing for the future through technology prediction has become an important factor for both countries and businesses. Existing methods of predicting through general expert advice are costly and time-consuming, and have limitations in that interests between groups and individuals can affect outcomes, and that they must rely entirely on subjective judgments based on individual experience.
On the other hand, when quantitative analysis is carried out using patent data that best reflects technical factors, cost and time can be relatively reduced and more objective results can be obtained unlike conventional methods. Research results can be used to establish a company's technology strategy and can also help government decision makers decide on support policies.