Search

During the Covid19 pandemic, many local businesses were struggling to survive. With the rising of click and collect solutions, business strives to survive by offering pick-up points and maintaining the social distance policy. The demand for goods and services of customers are met instantly from the local offline shops via digital platform. Meanwhile, local businesses in a specific geographical area are still striving from the hard times due to least popularity and competitivenes in the services such as delivery. Hyperlocal delivery comes into the logistics industry to meet the issues from local businesses. While the hyperlocal delivery terminology is still new, there have been some real world practices. This study attempts to define the hyperlocal delivery and explore multiple case studies to analyze the on-foot delivery as a rising of quick commerce in the gig economy. The multiple case studies will take place on the hyperlocal delivery business models in South Korea. The overview of the hyperlocal "on-foot" delivery will be described in four dimension of analysis; strategy, customer, cost-revenue, and process model. At the end, some research challenges and opportunities are presented to show how the research communities can contribute to the field. [ABSTRACT FROM AUTHOR]

Curriculum mining is a recent research area that applies a data-driven approach to assess students' learning behavior by discovering the curriculum model and compare it with the curriculum guideline. Some previous works exist to cope with the problem to discover the curriculum model from student database, by utilizing the concept of process mining. However, the challenges of discovering the curriculum model remain due to the different nature of student database from event log in twofold; the level of time granularity and variability of instance attributes. Previous works on curriculum mining that deal with conformance checking of the model between student's learning behavior and curriculum guideline are related to sequence matching alignment which is insufficient to understand the patterns of a group of students in a particular level of time granularity, i.e., Semester. This study proposes a curriculum mining methodology for curriculum assessment and students' learning behavior by creating segmentedtrace profiles. The segmented-trace profiles are extracted based on a local alignment of sequences and generated as the input for sequence matching alignment to assess whether the observed students' learning behavior match with the prior curriculum guideline. The profiles would be the features of the clustering approach. Real curriculum data has been used to test the effectivity of the methodology. The results show that the students can be grouped into various clusters per semester that have different characteristics for their learning behavior and performance. The results can be analyzed further to improve the curriculum guideline. [ABSTRACT FROM AUTHOR]