Dr. Teklu Urgessa - Academia.edu (original) (raw)
Uploads
Papers by Dr. Teklu Urgessa
Design Convergence Study, Oct 1, 2017
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017
Zenodo (CERN European Organization for Nuclear Research), Sep 29, 2023
Application of five implementations of three data mining classification techniques was experiment... more Application of five implementations of three data mining classification techniques was experimented for extracting important insights from tourism data. The aim was to find out the best performing algorithm among the compared ones for tourism knowledge discovery. Knowledge discovery process from data was used as a process model. 10-fold cross validation method is used for testing purpose. Various data preprocessing activities were performed to get the final dataset for model building. Classification models of the selected algorithms were built with different scenarios on the preprocessed dataset. The outperformed algorithm tourism dataset was Random Forest (76%) before applying information gain based attribute selection and J48 (C4.5) (75%) after selection of top relevant attributes to the class (target) attribute. In terms of time for model building, attribute selection improves the efficiency of all algorithms. Artificial Neural Network (multilayer perceptron) showed the highest i...
This paper presents an unobtrusive, low cost, light weight and easy to install laser sensor based... more This paper presents an unobtrusive, low cost, light weight and easy to install laser sensor based method for door people counting system. The system employed two paired laser sensors on two doors of one laboratory connected to Intel Edison board. The sensors have the capacity to detect people entering/exiting the doors. The Edison board computes the number of people based on sensing people and directions using the counting algorithm. The application for counting algorithm is developed using python programing language. The system provided fairly accurate counting results under different scenarios with the best scenario 95.3%. The scenarios tested were the sensors placement location on the door (knee (87%) or shoulder (93%) shoulder position using upload option to the cloud the improvement was 8% for changing location alone. Changing rules for uploading data from cloud to local server for the shoulder position has shown a 3% increase in accuracy (93% to 95.38%). The system has many implied applications like controlling room light, Controlling Heating, Ventilation and Air Conditioning (HVAC) and for tracking visitors' statistics in libraries, archival centers, and seat reservation systems.
The application of data mining techniques in identifying important patterns from different data-s... more The application of data mining techniques in identifying important patterns from different data-set has becoming astonishing. This academic research work is aimed at discovering the applicability of data mining techniques on ART data-set in discovering implicit and non trivial patterns that will give a clue to policy makers and the other searchers in the area. Classification and association data mining functionalities of the data mining techniques were tested on ART data taken from two hospitals in Ethiopia. The research proved that both the classification and association rule mining had revealed important patterns that triggers further study on the topic.
In this artificial intelligence time, speaker recognition is the most useful biometric recognitio... more In this artificial intelligence time, speaker recognition is the most useful biometric recognition technique. Security is a big issue that needs careful attention because of every activities have been becoming automated and internet based. For security purpose, unique features of authorized user are highly needed. Voice is one of the wonderful unique biometric features. So, developing speaker recognition based on scientific research is the most concerned issue. Nowadays, criminal activities are increasing day to day in different clever way. So, every country should have strengthen forensic investigation using such technologies. The study was done by inspiration of contextualizing this concept for our country. In this study, textindependent Amharic language speaker recognition model was developed using Mel-Frequency Cepstral Coefficients to extract features from preprocessed speech signals and Artificial Neural Network to model the feature vector obtained from the Mel-Frequency Cepst...
2021 IEEE International Conference on Technology, Research, and Innovation for Betterment of Society (TRIBES)
Nowadays, WSN is effectively using in healthcare application by means of body area network, which... more Nowadays, WSN is effectively using in healthcare application by means of body area network, which is senor-based wearable devices mainly utilized in health care prevention and monitoring applications. Limited resources make the WSN easily suffer from the problem of congestion, which degrade the performance of the network and hence reduced the reliability, which is most important requirement for WSN enabled healthcare. The objective of this paper is to design a routing protocol which can minimize communication delay and gain higher throughput with optimum network overhead. Proposed work present Load balancing technique toward Congestion minimization in WSN-enabled-Healthcare named LC-AOMDV mechanism. This mechanism utilizes calculation and regulation of queue length of intermediate nodes and reduced the data sending rate dynamically to avoid the congestion from the network. The proposed mechanism uses multi-hop multi-path routing and integrated into existing normal AOMDV routing protocol for efficient communications and more reliable health data delivery. The proposed work is simulated into Network Simulator 2.31 and results obtained are compared with results of existing routing technique. The results are analyzed by means of performance metrics like data/packet drops, PDR, routing overhead, throughput. Analysis shows that PDR is increased by around 4%, delay is decreased by 0.13 ms, and throughput is increased by 400 Kbps as compared to normal AOMDV. The proposed mechanism concluded and ensured that it gives better performance over existing technique and increase the reliability to transfer healthcare data.
This paper explores the scientific knowledge landscape at the intersection of user experience and... more This paper explores the scientific knowledge landscape at the intersection of user experience and human computer interaction field of study using keyword co-occurrence network analysis method. 499 journal articles from web of science from 1990 to 2016, which dealt with both user experience and human computer interaction were retrieved. The articles were portioned into three periods (1990-1999, 2000-2009, and 2010-2016). The author keywords network was constructed, clustered and overlaid by publication year and average citation of the articles from which the author keywords were extracted. Using the method it was possible to show various research directions in terms technological advances, methodological diversification and shifts in orientation with regard of study aspects of user experience over times.
Zenodo (CERN European Organization for Nuclear Research), Mar 10, 2023
International Journal of Advanced Computer Science and Applications
The insurance claim is a basic problem in insurance companies. Insurance insurers always have a c... more The insurance claim is a basic problem in insurance companies. Insurance insurers always have a challenge to the growing of insurance claim loss. Because there is the occurrence of claim fraud and the volume of claim data increases in the insurance companies. As a result, it is difficult to classify the insured claim status during the claim review process. Therefore, the aims of the study was to build a machine learning model that classifies and make motor insurance claim status prediction in machine learning approach. To achieve this study Missing value ratio, Z-Score, encoding techniques and entropy were used as data set preparation techniques. The final preprocessed data sets split using K-Fold cross validation techniques into training and testing sets. Finally the prediction model was built using Random Forest (RF) and Multi Class-Support Vector Machine (SVM).The performance of the models, RF and Multi-Class SVM classifiers were evaluated using Accuracy, Precision, Recall, and F-measure. The prediction accuracy of the model is capable of predicting the motor insurance claim status with 98.36% and 98.17% by RF and SVM classifiers respectively. As a result, RF classifier is slightly better than Multi-Class Support vector machines. Developing and implementing hybrid model to benefit from the advantages of different algorithms having graphical user interface to apply the solution to real world problem of the insurance company is a pressing future work.
Design Convergence Study, Oct 1, 2017
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017
Zenodo (CERN European Organization for Nuclear Research), Sep 29, 2023
Application of five implementations of three data mining classification techniques was experiment... more Application of five implementations of three data mining classification techniques was experimented for extracting important insights from tourism data. The aim was to find out the best performing algorithm among the compared ones for tourism knowledge discovery. Knowledge discovery process from data was used as a process model. 10-fold cross validation method is used for testing purpose. Various data preprocessing activities were performed to get the final dataset for model building. Classification models of the selected algorithms were built with different scenarios on the preprocessed dataset. The outperformed algorithm tourism dataset was Random Forest (76%) before applying information gain based attribute selection and J48 (C4.5) (75%) after selection of top relevant attributes to the class (target) attribute. In terms of time for model building, attribute selection improves the efficiency of all algorithms. Artificial Neural Network (multilayer perceptron) showed the highest i...
This paper presents an unobtrusive, low cost, light weight and easy to install laser sensor based... more This paper presents an unobtrusive, low cost, light weight and easy to install laser sensor based method for door people counting system. The system employed two paired laser sensors on two doors of one laboratory connected to Intel Edison board. The sensors have the capacity to detect people entering/exiting the doors. The Edison board computes the number of people based on sensing people and directions using the counting algorithm. The application for counting algorithm is developed using python programing language. The system provided fairly accurate counting results under different scenarios with the best scenario 95.3%. The scenarios tested were the sensors placement location on the door (knee (87%) or shoulder (93%) shoulder position using upload option to the cloud the improvement was 8% for changing location alone. Changing rules for uploading data from cloud to local server for the shoulder position has shown a 3% increase in accuracy (93% to 95.38%). The system has many implied applications like controlling room light, Controlling Heating, Ventilation and Air Conditioning (HVAC) and for tracking visitors' statistics in libraries, archival centers, and seat reservation systems.
The application of data mining techniques in identifying important patterns from different data-s... more The application of data mining techniques in identifying important patterns from different data-set has becoming astonishing. This academic research work is aimed at discovering the applicability of data mining techniques on ART data-set in discovering implicit and non trivial patterns that will give a clue to policy makers and the other searchers in the area. Classification and association data mining functionalities of the data mining techniques were tested on ART data taken from two hospitals in Ethiopia. The research proved that both the classification and association rule mining had revealed important patterns that triggers further study on the topic.
In this artificial intelligence time, speaker recognition is the most useful biometric recognitio... more In this artificial intelligence time, speaker recognition is the most useful biometric recognition technique. Security is a big issue that needs careful attention because of every activities have been becoming automated and internet based. For security purpose, unique features of authorized user are highly needed. Voice is one of the wonderful unique biometric features. So, developing speaker recognition based on scientific research is the most concerned issue. Nowadays, criminal activities are increasing day to day in different clever way. So, every country should have strengthen forensic investigation using such technologies. The study was done by inspiration of contextualizing this concept for our country. In this study, textindependent Amharic language speaker recognition model was developed using Mel-Frequency Cepstral Coefficients to extract features from preprocessed speech signals and Artificial Neural Network to model the feature vector obtained from the Mel-Frequency Cepst...
2021 IEEE International Conference on Technology, Research, and Innovation for Betterment of Society (TRIBES)
Nowadays, WSN is effectively using in healthcare application by means of body area network, which... more Nowadays, WSN is effectively using in healthcare application by means of body area network, which is senor-based wearable devices mainly utilized in health care prevention and monitoring applications. Limited resources make the WSN easily suffer from the problem of congestion, which degrade the performance of the network and hence reduced the reliability, which is most important requirement for WSN enabled healthcare. The objective of this paper is to design a routing protocol which can minimize communication delay and gain higher throughput with optimum network overhead. Proposed work present Load balancing technique toward Congestion minimization in WSN-enabled-Healthcare named LC-AOMDV mechanism. This mechanism utilizes calculation and regulation of queue length of intermediate nodes and reduced the data sending rate dynamically to avoid the congestion from the network. The proposed mechanism uses multi-hop multi-path routing and integrated into existing normal AOMDV routing protocol for efficient communications and more reliable health data delivery. The proposed work is simulated into Network Simulator 2.31 and results obtained are compared with results of existing routing technique. The results are analyzed by means of performance metrics like data/packet drops, PDR, routing overhead, throughput. Analysis shows that PDR is increased by around 4%, delay is decreased by 0.13 ms, and throughput is increased by 400 Kbps as compared to normal AOMDV. The proposed mechanism concluded and ensured that it gives better performance over existing technique and increase the reliability to transfer healthcare data.
This paper explores the scientific knowledge landscape at the intersection of user experience and... more This paper explores the scientific knowledge landscape at the intersection of user experience and human computer interaction field of study using keyword co-occurrence network analysis method. 499 journal articles from web of science from 1990 to 2016, which dealt with both user experience and human computer interaction were retrieved. The articles were portioned into three periods (1990-1999, 2000-2009, and 2010-2016). The author keywords network was constructed, clustered and overlaid by publication year and average citation of the articles from which the author keywords were extracted. Using the method it was possible to show various research directions in terms technological advances, methodological diversification and shifts in orientation with regard of study aspects of user experience over times.
Zenodo (CERN European Organization for Nuclear Research), Mar 10, 2023
International Journal of Advanced Computer Science and Applications
The insurance claim is a basic problem in insurance companies. Insurance insurers always have a c... more The insurance claim is a basic problem in insurance companies. Insurance insurers always have a challenge to the growing of insurance claim loss. Because there is the occurrence of claim fraud and the volume of claim data increases in the insurance companies. As a result, it is difficult to classify the insured claim status during the claim review process. Therefore, the aims of the study was to build a machine learning model that classifies and make motor insurance claim status prediction in machine learning approach. To achieve this study Missing value ratio, Z-Score, encoding techniques and entropy were used as data set preparation techniques. The final preprocessed data sets split using K-Fold cross validation techniques into training and testing sets. Finally the prediction model was built using Random Forest (RF) and Multi Class-Support Vector Machine (SVM).The performance of the models, RF and Multi-Class SVM classifiers were evaluated using Accuracy, Precision, Recall, and F-measure. The prediction accuracy of the model is capable of predicting the motor insurance claim status with 98.36% and 98.17% by RF and SVM classifiers respectively. As a result, RF classifier is slightly better than Multi-Class Support vector machines. Developing and implementing hybrid model to benefit from the advantages of different algorithms having graphical user interface to apply the solution to real world problem of the insurance company is a pressing future work.