MyEvents: A Personal Visual Analytics Approach for Mining Key Events and Knowledge Discovery in Support of Personal Reminiscence

Abstract

Reminiscence is an important aspect in our life. It preserves precious memories, allows us to form our own identities and encourages us to accept the past. Our work takes the advantage of modern sensor technologies to support reminiscence, enabling self‐monitoring of personal activities and individual movement in space and time on a daily basis. This paper presents MyEvents, a web‐based personal visual analytics platform designed for non‐computing experts, that allows for the collection of long‐term location and movement data and the generation of event mementos. Our research is focused on two prominent goals in event reminiscence: (1) selection subjectivity and human involvement in the process of self‐knowledge discovery and memento creation; and (2) the enhancement of event familiarity by presenting target events and their related information for optimal memory recall and reminiscence. A novel multi‐significance event ranking model is proposed to determine significant events in the personal history according to user preferences for event category, frequency and regularity. The evaluation results show that MyEvents effectively fulfils the reminiscence goals and tasks.

A Survey of Human‐Centered Evaluations in Human‐Centered Machine Learning

“…Although a high accuracy is often desired of machine learning models, the focus of many HCML papers is on explaining a model or creating models that respond to user preferences instead of emphasizing model quality. We thus found that only 18% of studies measured the accuracy of the model(s) (e.g., [EKC∗20; LXL∗18; PZDD19]). Of these, the measure of accuracy is often studied and derived from direct comparison to ground truth data [BAL∗15], or varied among study conditions [DLW∗17].…”

“…The predominant data types are multivariate data (e.g., [PNKC20; WMJ∗19; XMT∗20]); text (e.g., [ARO∗17; ESD∗19; LLL∗19]); and images (e.g., [CRH∗19; LSL∗17; SSSE20]). Only very few papers use other data types like videos (e.g., [KAY∗19]) or geo data (e.g., [PZDD19]).…”