Approximately three rhinos are poached daily in South Africa. Rhino poaching is a serious problem that a ects not only the rhino population of South Africa, but also the rhino population of the world. South Africa has the largest rhino population and of those rhinos the largest number can be found in the Kruger National Park (KNP). The KNP has been hit the hardest by the poaching epidemic, losing 1,175 rhinos in 2015 alone. Two big challenges are the size of the park and the unknown locations of both the poachers and new poaching events. The KNP is the size of a small country and there are simply not enough rangers to patrol this area e ectively. A costly solution would be to employ more rangers, but a proposed alternative is to reduce the search space and thus ensure that the rangers are allocated to the high risk areas first. A mathematical model was developed in the form of a Bayesian network (BN) to infer the most important factors contributing to poaching events and to model the rhino poaching problem. This model can be used to predict the area in which a future poaching attack could take place and thereby reduce the search area of rangers. The model also serves as a vehicle to enhance the understanding of the problem and encourage reasoning and discussion amongst decision makers. The map of the KNP is divided into cells and each cell is given a poaching probability, based on the outcome of the BN. This probability map forms a heat map that can be shown to the control centre and rangers can then be sent to the respective hotspots on the map. This is a proactive approach, which is in stark contrast to the numerous reactive approaches attempted thus far. This is the first BN modelling approach to the rhino poaching problem, and it is also the first BN application to wildlife crime. Previous applications of BN have only gone so far as environmental modelling, but not wildlife crimes. In this study the rhino poaching problem was shifted from a complex, ill-structured space to a position where researchers can begin to address the underlying problems by using a causal model as the vehicle for understanding the complex interplay between the factors a ecting poaching events. / Ongeveer drie renosters word daagliks in Suid-Afrika gestroop. Renosterstroping is 'n ernstige probleem wat nie net die renosterbevolking van Suid-Afrika raak nie, maar ook die res van die wêreld. Suid-Afrika het die grootste renoster bevolking in die wêreld, en die grootste getal van dié renosters word in die Kruger Nasionale Park (KNP) aangetref. Die KNP word die ergste geraak deur die stropings epidemie en 1,175 renosters is in 2015 gestroop. Twee groot uitdagings is die grootte van die park, asook die onbekende posisies van beide die stropers en die nuwe stropingsaanvalle. Die KNP is die grootte van 'n klein land en daar is eenvoudig nie genoeg veldwagters om hierdie area e ektief te patrolleer nie. 'n Duur oplossing sou wees om meer veldwagters aan te stel, maar 'n alternatief is om die soekarea van die veldwagters te verklein en sodoende te verseker dat die veldwagters die hoë-risiko areas eerste, en meer gereeld, patrolleer. 'n Wiskundige model in die vorm van 'n Bayesiese netwerk (BN) is ontwikkel om die belangrikste faktore te bepaal wat bydra tot stropingsaanvalle en uiteindelik die probleem te modelleer. Hierdie model kan gebruik word om die area te voorspel waar 'n stropingsaanval moontlik kan plaasvind en die soekarea van die veldwagters te verminder. Dit dien ook as 'n kanaal om die begrip van die probleem te verbeter en redenasie en bespreking onder besluitnemers aan te moedig. Die kaart van die KNP word in selle verdeel en aan elke sel word 'n stropingswaarskynlikheid toegeken gebaseer op die uitkoms van die BN. Hierdie waarskynlikheidskaart vorm 'n "hittekaart" wat aan die kontrolesentrum gewys kan word, en veldwagters kan dan na die onderskeie responskolle op die kaart gestuur word. Hierdie pro-aktiewe benadering is in teenstelling met huidige reaktiewe benaderings. Hierdie is die eerste BN modellering benadering tot die renosterstropingsprobleem, en dit is ook die eerste BN toepassing tot natuurlewe-misdaad. Vorige toepassings van BNs het omgewingsmodellering aangespreek, maar nie natuurlewe-misdade nie. In hierdie studie word aangetoon hoe die renosterstropings probleem geskuif is vanaf 'n komplekse, swak gestruktureerde probleemruimte na 'n omgewing waar navorsers kan begin om die onderliggende probleme aan te spreek deur gebruik te maak van 'n kausale model as die voertuig van begrip om die komplekse wisselwerking tussen faktore wat 'n stropingsaanval beïnvloed, te verstaan. / Thesis (PhD)--University of Pretoria, 2017. / Electrical, Electronic and Computer Engineering / PhD / Unrestricted

Family conversation between homes is difficult to initiate at mutually agreeable times as neither participant has exact knowledge of the other's activities or intentions. Whether calling to plan an important family gathering or simply to connect with family members, the question is: Is now a good time to call? People expect friends and family to learn their activity patterns and to minimize interruptions when calling. Can technology provide awareness cues to the caller, even prior to the initiation of the call? This research focuses on sampling the everyday activities of home life to determine environmental factors, which may serve as an indicator for availability. These external factors may be effective for identifying household routines of availability and useful in determining when to initiate conversation across homes. Several workplace studies have shown a person's interruptibility can be reliably assessed and modeled from specific environmental cues; this work looks for similar predictive power in the home. Copresence, location, and activity in the home were investigated as correlates to availability and for their effectiveness within the social protocol of family conversation. These studies indicate there are activities that can be sensed, either in real-time or over some time span, that correlate to self-reported availability. However, the type and amount of information shared is dependent upon individual preferences, social accessibility, and patterns of activities. This research shows friends and family can improve their predictions of when to call if provided additional context, and suggests that abstract representations of either routines or explicit availability status is sufficient and may be preferred by providers. Availability prediction is feasible in the home and useful to those outside the home, but the level of detail to provide in particular situations needs further study. This work has implications for the development of groupware systems, the automatic sensing of activity to deal with interruption, and activity recognition in the home.