Browsing by Author "Mekonnen, Fekadu"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item Application of Data Mining Techniques to Support Customer Relationship Management! (Crm) at Ethiopian Telecommunicate Tion Corporation (Etc(Addis Ababa University, 2004-06) Mekonnen, Fekadu; G/Yesus, Nega (PhD)The value of relevant and reliable information in this globalized. competitive and dynamic business environment is 100 high as it allows optimal decision making in all respects. Especially, in customer-oriented businesses like in the telecom industry, adequate in((ml1ation regarding Customers is villa so that appropriate customer relationships that maximize mutual benefit can be established) Behavioral based customer segmentation, ,which is one of the core applications of Customer Relationship Management (CRM). provides useful insights for designing and implementing an appropriate CHM strategies and programs that remit in success. To this end. data mining has strong potential in exploring natural segmentation schemes that lies within customers' data This study is an attempt to explore the customers' data to find underlying customer segments that may be use/ill Jar marketing decision making. The research followed the CRISP data mining process model. First the business problem was analyzed and a corresponding data mining tool, techniques and algorithms were selected. Besides, relevant data was collected, analyzed and prepared. Then, an automatic cluster detection models were bull to choose the best attributes for the final segmentation. Next, the final automatic cluster detection model was generated, analyzed and evaluated together with domain experts with different parameters. Finally, a decision tree model, selling cluster index as dependent variable, was built Corresponding rules to the decision tree were also generated. The results obtained were encouraging and can be further refined .for possible deploymentItem Informal Settlement and Building Construction Identification Using Change Detection: Case of Addis Ababa(Addis Ababa University, 2015-02) Mekonnen, Fekadu; Atnafu, Solomon (PhD)Population growths together with poverty are obstacles in the development and sustainable life. Lack of formal land provision and the high demand for housing increased the growth of the illegal/informal sector of housing. Getting up to date data of informal settlement from public register and urban cadaster are great challenge. This is because they are dynamic and often changing structures and forms. This is especially the case in developing countries. Informal settlement information is required for management and planning activities of urban region. The purpose of the study is to detect and analyze informal settlements using multi-temporal aerial photographs and cadastral map using change detection. The model employed unsupervised building identification and image differencing change detection in the identification of informal settlements and building construction. First, buildings are identified from the aerial photographs in the initial (1992) and target (2010) reference times separately. Then, the change detection between building maps in the initial and target reference times are used to identify the buildings constructed in the study period (1992-2010). Finally, the change detection between the labeled cadastral map and change detected building map are used to identify the informal settlements and building construction. The model is implemented using Matlab. The model was tested using sections of woredas 10 and 11 in Yeka sub-city, Addis Ababa. Manually identified informal buildings are used in the evaluation of performance of the system. Overall building identification and informal settlement and building construction identification accuracy were tested by Confusion Matrices. It has been found that tremendous informal settlements occurred over the study period. The results indicated that from the new settlement constructed between the periods 1992 to 2010 28.9% are informal settlements with annual growth rate around 1.61%. Keywords: Informal Settlement Identification, Building Identification, Image Differencing Change Detection, Cadastral Map, Multi-temporal Aerial Photograph, Addis Ababa CityItem Informal Settlement and Building Construction Identification Using Change Detection: Case of Addis Ababa(Addis Ababa University, 2015-02) Mekonnen, Fekadu; Atnafu, Solomon (PhD)Population growths together with poverty are obstacles in the development and sustainable life. Lack of formal land provision and the high demand for housing increased the growth of the illegal/informal sector of housing. Getting up to date data of informal settlement from public register and urban cadaster are great challenge. This is because they are dynamic and often changing structures and forms. This is especially the case in developing countries. Informal settlement information is required for management and planning activities of urban region. The purpose of the study is to detect and analyze informal settlements using multi-temporal aerial photographs and cadastral map using change detection. The model employed unsupervised building identification and image differencing change detection in the identification of informal settlements and building construction. First, buildings are identified from the aerial photographs in the initial (1992) and target (2010) reference times separately. Then, the change detection between building maps in the initial and target reference times are used to identify the buildings constructed in the study period (1992-2010). Finally, the change detection between the labeled cadastral map and change detected building map are used to identify the informal settlements and building construction. The model is implemented using Matlab. The model was tested using sections of woredas 10 and 11 in Yeka sub-city, Addis Ababa. Manually identified informal buildings are used in the evaluation of performance of the system. Overall building identification and informal settlement and building construction identification accuracy were tested by Confusion Matrices. It has been found that tremendous informal settlements occurred over the study period. The results indicated that from the new settlement constructed between the periods 1992 to 2010 28.9% are informal settlements with annual growth rate around 1.61%. Keywords: Informal Settlement Identification, Building Identification, Image Differencing Change Detection, Cadastral Map, Multi-temporal Aerial Photograph, Addis Ababa City