Article Preview
Top1. Introduction
An average user has many activities in a typical day. Most of these activities are performed on a daily basis. Due to the stress at work and other factors, people tend to forget some activities that need to be taken care of (Rokhman & Saifuddin 2016). In general, these activities can be broadly classified into two categories: time-based and location-based. For a time-based activity, it is expected that the activity is carried out at a specific time. For example, waking up at 7 a.m. every day is a time-based activity. In order not to forget these types of activities, a reminder can be set in the reminder system of the personal smart phone to alert the user. On the other hand, a location-based activity is performed at a specific location (Sohn et al. 2005), such as going to the post office which is close to your home. In some cases, it is inappropriate to remind the user about a location-based reminder because the office hours have passed and the location is closed. This problem can be solved by combining the time parameter with location. With the GPS sensors available on the smart phones, more and more recent location-based reminder (LBR) applications have incorporated both time and location conditions (Battin & Markande 2016, Lin & Hung 2014, Pamulapati & Li 2017).
Location-based services (LBS) have been used in a variety of applications, such as healthcare (Kooij et al. 2017, Sebastian-Viana et al. 2016), sports (Fister & Fister Jr. 2011), community (Sasao et al. 2017), and personal life (Gaurav & Shrivastava 2015, Giunchiglia et al. 2017, Oh et al. 2015), etc. Kooij et al. (2017) studied the reasons for non-adherence in the setting of automated reminders to improve guideline adherence on prescribing and administering medicines. Fister and Fister Jr. (2017) developed a drafting detection system in the sports Ironmans by using mobile devices to calculate the distance and time between the i-th and the (i+1)-th competitors. In (Sasao et al. 2017), a context-aware reminder application is examined on how local communities can give advice, tips, and mobile crowdsourcing requests to support their members. Oh et al. (2015) employ the Naïve Bayesian approach to predict user events based on the personalized latent features from heterogeneous data streams.
As an example of the location-based services, the currently existing location-based reminders do not consider the user’s preferences that might change on a daily basis (Deva et al. 2015, Ertugrul & Onal 2014, Nate et al. 2016, Pamulapati & Li 2017, Rokhman & Saifuddin 2016). In (Rokhman & Saifuddin 2016), although the visited locations are considered as user preferences, the reminder system is not able to recommend the best locations based on the user’s preferences and demographic information. Deva et al. (2015) present a privacy-aware framework for proactive LBS that enables user-configured and context-sensitive privacy preferences, in form of rules, which are manually created, modified, and deleted by the user. SmartNotify presented in this paper is most close to iDoRemind (Pamulapati & Li 2017) in many ways such that both reminder applications allow the user to share the reminders on social media, create groups, and suggest locations to visit, but the locations recommended by iDoRemind are based on the customer ratings and are different from the locations recommended by the personalized data mining approach proposed in this paper.