Million Meshesha

A Corpus for Amharic-English Speech Translation: The Case of Tourism Domain

Speech translation research for the major languages like English, Japanese and Spanish has been conducted since the 1980’s. But no attempt were made in speech translation to/from the under-resourced language like Amharic. These activities suffered from the lack of Amharic speech and Amharic-English text corpus suited for the development of speech translation between the two languages. In this paper, therefore, an attempt has been made to collect, translate and record speech data from resourced language (English) to under-resourced language (Amharic) taking a Basic Traveler Expression Corpus (BTEC) as domain. Since there is no any Amharic text and speech corpus readily available for speech translation purposes, first, 7.43 h of Amharic read-speech has been prepared from 8,112 sentences, and second, 19,972 parallel Amharic-English corpus has been prepared taking tourism as an application domain. The Amharic speech data is recorded using smart-phone based application tool, LIG-Aikuma under a normal working environment. With the availability of such standard speech and text corpus, researcher will find a ground to further explore speech translation to/from under resourced languages.

Developing a Knowledge-Based System for Diagnosis and Treatment of Malaria

Malaria is a serious and fatal disease caused by a parasite that can infect a certain type of mosquito which feeds on human blood. It is a public health problem in Ethiopia and a major cause of illness and death. More than 75% of the total land of Ethiopia is malarious affecting more than 68% of the population, making malaria the leading public health problem in Ethiopia. In an effort to address such problems, it is important to develop knowledge-based system (KBS) that can provide advice for health professionals and patients to facilitate diagnosis and treatment of malaria patients. Experimental research design was used to developed prototype system. Purposive sampling technique was used to select domain experts for knowledge acquisition. The domain experts are selected from Jimma special hospital, Adama hospital and Agaro health centre. The knowledge was acquired using both structured and unstructured interviews from domain experts and represented by production rule, (if- then method). The users acceptance of the prototype system by visual interaction method that by showing the prototype system to the domain experts was conducted result is 83.21%. In addition, performance of the prototype system was evaluated using case testing method and produce result of 82.3%. It is promising to save the life of people in rural area where there is scarcity of health professionals and apparatus. In addition, it is possible to reduce time and cost of diagnosis and treatment in health centre by implementing intelligent systems. Developing in local languages, good interface programming language and in other techniques are the future works of the study.

Developing Knowledge-Based Systems Using Data Mining Techniques for Advising Secondary School Students in Field of Interest Selection

Ethiopia gives a highly emphasis on a secondary school and reform program of impressive expansion. In doing this, students interest towards the fields they are assigned to needs to be taken into consideration. This has been put into practice when the Ministry of Education and preparatory schools have assigned students in fields of studies based on their performance at secondary schools. However, they used only the students grade 10 Ethiopian General School Leaving Certificate Examination result to assign them. The objective of this study is to develop a knowledge-based systems using machine learning (data mining) techniques that consults the students in their field of study selection process. In this study, the hybrid model that was developed for academic research is used. To build the predictive model, 9364 sample students data from selected secondary schools are used. The sample data is preprocessed for missing values, outliers, noisy and errors. Then the model is experimented using decision tree (j48) and rule induction (PART) algorithms. In this study as compared to j48, the PART unpruned decision list algorithm has 98.003% predictive performance. Thus, the knowledge discovered with this algorithm is further used to build the knowledge-based systems. Hence, the Java program is used to integrate data mining results to knowledge-based systems. As a result, the developed knowledge based-systems is used to predict students field of study based on their performance at secondary school. The study concludes that, to build the accurate knowledge-based systems discovering knowledge using data mining techniques is significant.

Experimenting Statistical Machine Translation for Ethiopic Semitic Languages: The Case of Amharic-Tigrigna

In this research an attempt have been made to experiment on Amharic-Tigrigna machine translation for promoting information sharing. Since there is no Amharic-Tigrigna parallel text corpus, we prepared a parallel text corpus for Amharic-Tigrigna machine translation system from religious domain specifically from bible. Consequently, the data preparation involves sentence alignment, sentence splitting, tokenization, normalization of Amharic-Tigrigna parallel corpora and then splitting the dataset into training, tuning and testing data. Then, Amharic-Tigrigna translation model have been constructed using training data and further tuned for better translation. Finally, given target language model, the Amharic-Tigrigna translation system generates a target output with reference to translation model using word and morpheme as a unit. The result we found from the experiment is promising to design Amharic-Tigrigna machine translation system between resource deficient languages. We are now working on post-editing to enhance the performance of the bi-lingual Amharic-Tigrigna translator.

Towards enabling tacit knowledge externalization using mobile phone: The case of participatory agricultural innovation in Ethiopia

Mobile technology has huge potential to address challenges in the areas of tacit knowledge elicitation, capturing, sharing and integration in participatory agricultural innovation where farming communities are actively involving. Past studies emphasized the behavioral and managerial perspectives of tacit knowledge and limited studies have been conducted from ICT perspective. Those ICT related studies focused more on advanced technologies including Web 2.0, groupware, audio & videoconferencing tools which are beyond the reach of the rural communities. The purpose of this study is therefore, to identify various mechanisms of tacit knowledge elicitation and sharing in participatory agricultural research process and explore how mobile based application can enhance the process. The output of the study will contribute to the development of mobile-enabled communication system that can further enhance participatory agricultural innovation through facilitating articulation, capturing, sharing and integration of tacit knowledge of multiple stakeholders including farmers.

Text to speech synthesis for ethiopian semitic languages: Issues and the way forward

Currently there is a tremendous increase in electronic text in various languages, including Ethiopian Semitic languages. Yet, it is inaccessible to visually impaired people and the illiterate. To come up with high quality text to speech synthesizers for local languages, it is imperative that research in natural language processing and synthetic speech generation be improved. Accordingly, we critically reviewed core issues in text to speech synthesis for Ethiopian Semitic languages and revealed that further research to improve the quality of text to speech synthesis for Ethiopian Semitic languages is mandatory. To optimize linguistic resources required for high quality synthetic speech output we propose designing a generic bilingual Text to Speech framework for Ethiopian Semitic languages.

Barriers of Knowledge Sharing Among Employees: The Case of Commercial Bank of Ethiopia

Commercial banks are one of the main engines that enhance the economic growth of the country by managing financial transactions. Banks process and use information to run their business. Knowledge is one of the strategic resources that commercial banks use to increase their internal efficiency and to operate competitively. Knowledge-sharing barriers hinder the smooth flow of knowledge among employees which often results in negative consequences such as customer dissatisfaction, low employee learning and poor service quality. This research identified complex individual, organisational and technological factors that affect knowledge sharing and puts forward interventions that can improve the culture of knowledge sharing in an organisation. The research also revealed that although organisations put much emphasis on the development of a technological infrastructure as a means to develop their knowledge management, it is the organisational and individual factors that may prove to be more important in improving organisational knowledge management. This research has a theoretical contribution for the generalisability of existing knowledge sharing theory across different socioeconomic contexts, in particular in Ethiopia.

Enhancing semantic relation quality of UMLS knowledge sources

The quality of semantic tuples (semantic triples forming subject-predicate-object) has significant impact in most text mining and knowledge discovery applications. The practical success and usability of these applications momentously depends on the quality of the extracted semantic triples. Most biomedical semantic resources have been developed for different contexts focusing on the structural representation but with less attention on the acceptability and naturalness of the individual semantic triples. In this article, we presented an integrated approach for enhancing the quality of semantic tuples in the UMLS knowledge sources. The approach is based on the integration of three existing auditing techniques: avoiding redundant classifications of semantic concepts, reducing hierarchical and associative relationship inconsistencies. We evaluated the approach based on the number of identified wrongly assigned concepts and inconsistent relationships obtained. The quality of each semantic triple is evaluated based on the acceptability and naturalness of the semantic tuples. The evaluation shows promising results. In the evaluation, we have extracted 10,082 semantic triples randomly from UMLS and obtained 5646 taxonomically and 4436 non-taxonomically related semantic triples. 826 concepts are found redundantly classified and 352 are found hierarchically inconsistent. In non-taxonomic semantic triples, out of 4436, 726 are found to be inconsistent. The quality (acceptability and naturalness) of each semantic triples of the first 100 are also evaluated using domain experts. The Cohens kappa coefficient is used to measure the degree of agreement between the annotators and the result is promising (0.8).