Browsing by Author "Gebeyehu, Zewdu"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    A frame work for Multi Agent Referralde Clslon Suppor
    (Addis Ababa University, 2000-07) Gebeyehu, Zewdu; Libsie, Mulugeta(PhD)
    Medical Referral Systems aim at achieving high standards of care by improving patient outcomes and decreasing costs through optimal use of medical services. The success of these systems is highly dependent on the quality of referral decisions. A referral decision is a clinical decision by which physicians determine referral indication, type of required services and selection of appropriate providers. This inherently complicated process depends on a complex mix of patient, provider and healthcare system determinants. It requires medical and non-medical knowledge that exist distributed among several healthcare providers. In addition, the process requires communication and coordination of information between providers. Little research is done in studying how to provide an intelligent decision support that can help physicians in making better referral decisions. Most of the works focus in studying how to improve referral communication by employing e-Referral systems. This in turn leads to minimal effect in terms of achieving the objectives of Medical Referral Systems. This thesis proposes a Multi-Agent Referral Decision Support (MARDS) framework aimed at improving quality of referral decisions. It specifically presents a Multi-Criteria Provider Selection (MCPS) model that can be employed by the MARDS to aid physicians in making a better provider selection. The core part of the selection model is simulated by a Multi-Agent Provider Selection System (MPSS), which is developed using JADE (Java Agent DEvelopment Framework). The system is experimented with limited but relevant data and has shown the feasibility of the selection strategy. The result of this work is believed to be one step towards enhancing existing e-Referral systems with intelligent referral decision support. Moreover, there is a possibility of employing the decision support framework to other application areas of e-Health (like e-Consultation) with some modifications. Keywords: Medical Referral Systems, Referral Decisions, e-Referral Systems, Multi-Agent Systems, Decision Support Systems, Distributed Systemsviii
  • No Thumbnail Available
    Item
    A frame work for Multi Agent Referralde Clslon Suppor
    (Addis Ababa University, 2000-07) Gebeyehu, Zewdu; Libsie, Mulugeta(PhD)
    Medical Referral Systems aim at achieving high standards of care by improving patient outcomes and decreasing costs through optimal use of medical services. The success of these systems is highly dependent on the quality of referral decisions. A referral decision is a clinical decision by which physicians determine referral indication, type of required services and selection of appropriate providers. This inherently complicated process depends on a complex mix of patient, provider and healthcare system determinants. It requires medical and non-medical knowledge that exist distributed among several healthcare providers. In addition, the process requires communication and coordination of information between providers. Little research is done in studying how to provide an intelligent decision support that can help physicians in making better referral decisions. Most of the works focus in studying how to improve referral communication by employing e-Referral systems. This in turn leads to minimal effect in terms of achieving the objectives of Medical Referral Systems. This thesis proposes a Multi-Agent Referral Decision Support (MARDS) framework aimed at improving quality of referral decisions. It specifically presents a Multi-Criteria Provider Selection (MCPS) model that can be employed by the MARDS to aid physicians in making a better provider selection. The core part of the selection model is simulated by a Multi-Agent Provider Selection System (MPSS), which is developed using JADE (Java Agent DEvelopment Framework). The system is experimented with limited but relevant data and has shown the feasibility of the selection strategy. The result of this work is believed to be one step towards enhancing existing e-Referral systems with intelligent referral decision support. Moreover, there is a possibility of employing the decision support framework to other application areas of e-Health (like e-Consultation) with some modifications. Keywords: Medical Referral Systems, Referral Decisions, e-Referral Systems, Multi-Agent Systems, Decision Support Systems, Distributed Systemsviii
  • No Thumbnail Available
    Item
    Synthesis and Characterization of Some Nitridc Molybdenum and Tungsten Complexes
    (Addis Ababa Universty, 1993-06) Gebeyehu, Zewdu; Sey.forth, J. (PhD)
    The nitridomolybdenum(VI) complexes MoI'TOI (OPPh^)p and LMoUCl^j were synthesized from lioCl^CCH^CN)^, \ sodium azid, triphenylphosphine oxide and totrabutylaromonium chloride respectively in acetonitrile. The reaction of fioClr with sodium azide and triphenylphosphine oxide also gave i'ioNCl^ (OPPh^^j which is found to be anew way of preparing nitridomolybdenum complexes. The nitridomolybdenum complexes obtained were analytically pure. Attempts to synthesize nitridotungsten(VI) complexes by reaction of wCl^ with sodium azide in acetonitrile gave impure WNC1^(CH^ON)^ complex. Reaction of the recrystallized and decomposition product of WNCl^(CHyOll)p with triphenylphosphine oxide in methylenechloride gave pure VfifCl^(0PPh^)2 in low yield. Similar reaction of ViWCl^(CH^CN)p with tetrabutylammonium chloride an.d 2,2'-dipyridyl in acetonitrile gave only impure products