Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845565
Silma Koekemoer, Rossouw von Solms
Smart cities have become a fashionable trend and created much excitement internationally, as well as in South Africa. However, the developing countries of the South can ill-afford the required investment, and mostly do not have the skills or the infrastructure to achieve this status. This paper employs a literature review to identify key elements which are pervasive in smart cities, and packages these into a localised maturity assessment model for South African municipalities, to self-assess their maturity on the continuum towards achieving smart status. The value of the maturity model, however, is not the model or the baseline assessment, but the information that is consolidated and informs an achievable roadmap for every municipality, from metropolitan to small rural administrations, to embark on a customised smartification journey, towards smarter city status within their own context.
{"title":"A Smart City Maturity Assessment Model for South African Municipalities","authors":"Silma Koekemoer, Rossouw von Solms","doi":"10.23919/IST-Africa56635.2022.9845565","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845565","url":null,"abstract":"Smart cities have become a fashionable trend and created much excitement internationally, as well as in South Africa. However, the developing countries of the South can ill-afford the required investment, and mostly do not have the skills or the infrastructure to achieve this status. This paper employs a literature review to identify key elements which are pervasive in smart cities, and packages these into a localised maturity assessment model for South African municipalities, to self-assess their maturity on the continuum towards achieving smart status. The value of the maturity model, however, is not the model or the baseline assessment, but the information that is consolidated and informs an achievable roadmap for every municipality, from metropolitan to small rural administrations, to embark on a customised smartification journey, towards smarter city status within their own context.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"51 8","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113936276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845552
Francis Katende, J. Katende
In this article, Simulink is used to implement the open and short circuit tests of three-phase transformers. With most universities switching to online methods of teaching in the wake of the coronavirus pandemic, this article outlines how to perform the transformer tests for undergraduate electric machines classes using Simulink. The aim is present the virtual experiments whose results lead to determination of the approximate per-phase equivalent circuit model of a three-phase power transformer. The article will be of help for students who cannot attend conventional practical laboratory sessions due to COVID-19 standard operating procedures restrictions or shortage of equipment and will be a great supplement for those who can.
{"title":"Determination of Per-Phase Equivalent Circuit Parameters of Three-Phase Transformer Using MATLAB/Simulink","authors":"Francis Katende, J. Katende","doi":"10.23919/IST-Africa56635.2022.9845552","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845552","url":null,"abstract":"In this article, Simulink is used to implement the open and short circuit tests of three-phase transformers. With most universities switching to online methods of teaching in the wake of the coronavirus pandemic, this article outlines how to perform the transformer tests for undergraduate electric machines classes using Simulink. The aim is present the virtual experiments whose results lead to determination of the approximate per-phase equivalent circuit model of a three-phase power transformer. The article will be of help for students who cannot attend conventional practical laboratory sessions due to COVID-19 standard operating procedures restrictions or shortage of equipment and will be a great supplement for those who can.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125125373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845638
R. Koubaa, Sana Rekik, M. Jmaiel, Million Tesfaye, Maame Esi Amekudzi, Muyepa Anthony, Moges Asressie, Konstantinos Antypas
Strategic partnerships are very important for the successful deployment of e-health as they play a crucial role in achieving common goals and creating an added value for the involved partners. In this paper, we will provide relevant information about strategic partnerships in e-health deployment in four African countries, namely Ethiopia, Ghana, Malawi, and Tunisia. A Partnership Assessment Tool is developed to analyze different aspects of partnerships and classify them. According to the analysis, 11 partnerships were strategic amongst the 15 identified. Findings analysis also shows that certain aspects, mainly sustainability, have to be enhanced to guarantee the impact of partnerships after the ending of its actions. Increased governmental support is required in addition to international funding resources to the successful deployment of e-health in the participating countries.
{"title":"Strategic Partnerships in e-Health in Low and Lower Middle-Income Countries in Africa","authors":"R. Koubaa, Sana Rekik, M. Jmaiel, Million Tesfaye, Maame Esi Amekudzi, Muyepa Anthony, Moges Asressie, Konstantinos Antypas","doi":"10.23919/IST-Africa56635.2022.9845638","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845638","url":null,"abstract":"Strategic partnerships are very important for the successful deployment of e-health as they play a crucial role in achieving common goals and creating an added value for the involved partners. In this paper, we will provide relevant information about strategic partnerships in e-health deployment in four African countries, namely Ethiopia, Ghana, Malawi, and Tunisia. A Partnership Assessment Tool is developed to analyze different aspects of partnerships and classify them. According to the analysis, 11 partnerships were strategic amongst the 15 identified. Findings analysis also shows that certain aspects, mainly sustainability, have to be enhanced to guarantee the impact of partnerships after the ending of its actions. Increased governmental support is required in addition to international funding resources to the successful deployment of e-health in the participating countries.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125562310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845597
C. Umezuruike, C. Diji, John Patrick Asiimwe, H. Ngugi
Information systems have played vital roles in enhancing productivity in different domains using available data, especially in agriculture. With enormous data harvested from the direct farmers and remote sensor networks, there is still the gap of proper processing, and use to improve agricultural production in sub-Saharan Africa. As a means to an end, this work proposes a data product model that will boost agricultural production in sub-Saharan Africa. An object-oriented approach was employed to study the fundamental components of a data product and the processes involved. The study climaxed with the design of an implementable data product model using secondary data collected from secondary sources as input. The result was an interactive model that identified the actors, actions, sequence of data flow (Use Case and Sequence Diagram), and the model component. The proposed model has five distinct components: data production component; database component; Extract, Transform and Load component; the Data Team; and the Data Mart. The processes that go on within the components were itemized to be 7 processes.
{"title":"Data Product Model for boosting Agricultural Productivity in Sub-Saharan Africa","authors":"C. Umezuruike, C. Diji, John Patrick Asiimwe, H. Ngugi","doi":"10.23919/IST-Africa56635.2022.9845597","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845597","url":null,"abstract":"Information systems have played vital roles in enhancing productivity in different domains using available data, especially in agriculture. With enormous data harvested from the direct farmers and remote sensor networks, there is still the gap of proper processing, and use to improve agricultural production in sub-Saharan Africa. As a means to an end, this work proposes a data product model that will boost agricultural production in sub-Saharan Africa. An object-oriented approach was employed to study the fundamental components of a data product and the processes involved. The study climaxed with the design of an implementable data product model using secondary data collected from secondary sources as input. The result was an interactive model that identified the actors, actions, sequence of data flow (Use Case and Sequence Diagram), and the model component. The proposed model has five distinct components: data production component; database component; Extract, Transform and Load component; the Data Team; and the Data Mart. The processes that go on within the components were itemized to be 7 processes.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122619213","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845510
E. Micheni, Jackson Machii, Julius Murumba
Agriculture is undergoing a digital transformation because of population growth, climate change, and food security concerns. Agriculture is influenced by information technology in terms of cost reduction, efficiency, and sustainability. Precision agriculture employs IoT, deep learning, predictive analytics, and AI-based technologies to aid in the detection of plant diseases, pests, and poor plant nutrition in the field. The study’s objectives are as follows: 1) evaluate the role of smart technologies and their impact on precision agriculture sustainability; 2) assess the typical application of IoT data analytics and deep learning in precision agriculture; and 3) investigate the barriers to the adoption of sustainable precision farming. IoT technologies collect data and relay it to data analytics and deep learning for in-depth analysis. The findings indicate that data assists farmers in managing crop variety, phenotypes and selection, crop performance, soil quality, pH level, irrigation, and fertilizer application quantity. The study looks at typical application areas and critical success factors for precision agriculture. Technological issues, safety, privacy, cost, and legal issues influence the adoption of these technologies. Individual farmers, government, academics, and agricultural authorities will all benefit from the research. The study recommends the adoption and optimization of innovations and technologies e.g. mobile devices, access to better internet speed, low-cost and dependable satellites for positioning and imagery, and precision agriculture-optimized agricultural machinery. Future research should focus on the application of appropriate decision-support systems for implementing precision decisions.
{"title":"Internet of Things, Big Data Analytics, and Deep Learning for Sustainable Precision Agriculture","authors":"E. Micheni, Jackson Machii, Julius Murumba","doi":"10.23919/IST-Africa56635.2022.9845510","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845510","url":null,"abstract":"Agriculture is undergoing a digital transformation because of population growth, climate change, and food security concerns. Agriculture is influenced by information technology in terms of cost reduction, efficiency, and sustainability. Precision agriculture employs IoT, deep learning, predictive analytics, and AI-based technologies to aid in the detection of plant diseases, pests, and poor plant nutrition in the field. The study’s objectives are as follows: 1) evaluate the role of smart technologies and their impact on precision agriculture sustainability; 2) assess the typical application of IoT data analytics and deep learning in precision agriculture; and 3) investigate the barriers to the adoption of sustainable precision farming. IoT technologies collect data and relay it to data analytics and deep learning for in-depth analysis. The findings indicate that data assists farmers in managing crop variety, phenotypes and selection, crop performance, soil quality, pH level, irrigation, and fertilizer application quantity. The study looks at typical application areas and critical success factors for precision agriculture. Technological issues, safety, privacy, cost, and legal issues influence the adoption of these technologies. Individual farmers, government, academics, and agricultural authorities will all benefit from the research. The study recommends the adoption and optimization of innovations and technologies e.g. mobile devices, access to better internet speed, low-cost and dependable satellites for positioning and imagery, and precision agriculture-optimized agricultural machinery. Future research should focus on the application of appropriate decision-support systems for implementing precision decisions.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133653501","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845617
Onneile Bashapi, Thotobolo Morapedi
The current digital financial services ecosystem in Botswana does not have policies facilitating an open banking environment, i.e., lack of public or Open APIs, leading to the financial technology infrastructure not being interoperable. This paper aims to develop a conceptual framework that will enable the development of an interoperable financial ecosystem through collaborative open innovation in Botswana. A literature review was used to identify developments towards open banking or open APIs in other countries. The National ICT Policy Review and E-commerce Strategy for Botswana were reviewed to identify developments in Botswana regarding the financial technology infrastructure. The literature review showed the need for collaboration between banks and fintechs as they can both benefit from that relationship. The results of the study achieved all the objectives whereby (i) an Open API conceptual framework that will enable the development of an interoperable financial ecosystem through collaborative Open Innovation was developed, (ii) a mobile payment architecture based on the Open API conceptual framework and (iii) a mobile application as a proof-of-concept based on the developed Open API conceptual framework. Business benefits of the study were discussed.
{"title":"Collaborative Open Innovation through an Open API Approach to Enable the Development of Financial Technology Infrastructure: Case of Botswana","authors":"Onneile Bashapi, Thotobolo Morapedi","doi":"10.23919/IST-Africa56635.2022.9845617","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845617","url":null,"abstract":"The current digital financial services ecosystem in Botswana does not have policies facilitating an open banking environment, i.e., lack of public or Open APIs, leading to the financial technology infrastructure not being interoperable. This paper aims to develop a conceptual framework that will enable the development of an interoperable financial ecosystem through collaborative open innovation in Botswana. A literature review was used to identify developments towards open banking or open APIs in other countries. The National ICT Policy Review and E-commerce Strategy for Botswana were reviewed to identify developments in Botswana regarding the financial technology infrastructure. The literature review showed the need for collaboration between banks and fintechs as they can both benefit from that relationship. The results of the study achieved all the objectives whereby (i) an Open API conceptual framework that will enable the development of an interoperable financial ecosystem through collaborative Open Innovation was developed, (ii) a mobile payment architecture based on the Open API conceptual framework and (iii) a mobile application as a proof-of-concept based on the developed Open API conceptual framework. Business benefits of the study were discussed.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"129 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123519131","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845587
Mutshiene Deogratias Ekwanzala, M. Momba
Environment-related infections are a significant cause of mortality in developing countries and cause substantial morbidity in developed nations. This situation is exacerbated with the recent emergence of multi-and pan-drug resistant bacteria, where their spread and development are now a genuine health issue that is rapidly expanding worldwide. There is, therefore, a need for an integrated and curated database at a genomic level that will enable the tracking and identification of sources of resistant environmental to clinical infections for epidemiological containment. This paper presents GenoTrack - a genomic epidemiology web application that geospatially maps clinical bacterial genomes isolated from environmental and clinical settings focusing on their antibiotic resistance genes content. GenoTrack enables the geospatial mapping of genomes of selected bacteria using the cartographic display to yield a precise distribution of genomes in different settings. A phylogenomic tree showing the relatedness of analysed genomes and genome alignment features allows structural genomic element analysis. Here, we outline the development of GenoTrack and its implemented features, and we demonstrate its application to track genomes from different environments.
{"title":"GenoTrack, a Webtool to Geospatially Link Bacterial Genomes from Environmental and Clinical Settings","authors":"Mutshiene Deogratias Ekwanzala, M. Momba","doi":"10.23919/IST-Africa56635.2022.9845587","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845587","url":null,"abstract":"Environment-related infections are a significant cause of mortality in developing countries and cause substantial morbidity in developed nations. This situation is exacerbated with the recent emergence of multi-and pan-drug resistant bacteria, where their spread and development are now a genuine health issue that is rapidly expanding worldwide. There is, therefore, a need for an integrated and curated database at a genomic level that will enable the tracking and identification of sources of resistant environmental to clinical infections for epidemiological containment. This paper presents GenoTrack - a genomic epidemiology web application that geospatially maps clinical bacterial genomes isolated from environmental and clinical settings focusing on their antibiotic resistance genes content. GenoTrack enables the geospatial mapping of genomes of selected bacteria using the cartographic display to yield a precise distribution of genomes in different settings. A phylogenomic tree showing the relatedness of analysed genomes and genome alignment features allows structural genomic element analysis. Here, we outline the development of GenoTrack and its implemented features, and we demonstrate its application to track genomes from different environments.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124950505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845635
Mohammad Shoaib Ibne Saleem Casseem, S. Venkannah, Y. Bissessur
Recent problems in the world have highlighted the disadvantages of being a global village. Many countries have become over dependent on external sources for many basic commodities affecting the local primary sector. Food security is a major concern to small islands states like Mauritius and one major issue is the high cost of production and labor scarcity. Artificial intelligence can now be used to support the local entrepreneurs in their businesses, but the major problem is barrier to the introduction of new technologies due to lack of technical support to the local entrepreneurs. This project aims at the design of a system that is capable of harvesting tomatoes indoor in an autonomous way for an entrepreneur involved in Mauritius, who is currently facing various problems related mostly to a shortage of labour. The system was designed specifically for the company taking into consideration its requirements and constraints. The proposed system was a 3-axis robotic arm mounted on an Automated Guided Vehicle (AGV) capable of picking tomatoes using computer vision for identification and recognition. The spatial location of the fruits was obtained by means of stereovision, which is a technique consisting of two cameras viewing the scene from two different positions and then through triangulation, the real-world coordinates of the tomatoes were extracted. Using this information, the robotic arm was able to pluck and store the tomatoes. The AGV, on the other hand, was used to transport the robotic arm throughout the greenhouse and line following was employed so that the vehicle achieved an autonomous behaviour. The time for the robotic arm to harvest and store one tomato was approximately ten seconds, but this slow speed was compensated by the system’s ability to work for four hours straight and multiple shifts after charging.
{"title":"Design of a Tomato Harvesting Robot for Agricultural Small and Medium Enterprises (SMEs)","authors":"Mohammad Shoaib Ibne Saleem Casseem, S. Venkannah, Y. Bissessur","doi":"10.23919/IST-Africa56635.2022.9845635","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845635","url":null,"abstract":"Recent problems in the world have highlighted the disadvantages of being a global village. Many countries have become over dependent on external sources for many basic commodities affecting the local primary sector. Food security is a major concern to small islands states like Mauritius and one major issue is the high cost of production and labor scarcity. Artificial intelligence can now be used to support the local entrepreneurs in their businesses, but the major problem is barrier to the introduction of new technologies due to lack of technical support to the local entrepreneurs. This project aims at the design of a system that is capable of harvesting tomatoes indoor in an autonomous way for an entrepreneur involved in Mauritius, who is currently facing various problems related mostly to a shortage of labour. The system was designed specifically for the company taking into consideration its requirements and constraints. The proposed system was a 3-axis robotic arm mounted on an Automated Guided Vehicle (AGV) capable of picking tomatoes using computer vision for identification and recognition. The spatial location of the fruits was obtained by means of stereovision, which is a technique consisting of two cameras viewing the scene from two different positions and then through triangulation, the real-world coordinates of the tomatoes were extracted. Using this information, the robotic arm was able to pluck and store the tomatoes. The AGV, on the other hand, was used to transport the robotic arm throughout the greenhouse and line following was employed so that the vehicle achieved an autonomous behaviour. The time for the robotic arm to harvest and store one tomato was approximately ten seconds, but this slow speed was compensated by the system’s ability to work for four hours straight and multiple shifts after charging.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"172 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114422643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845545
E. O. Aliyu, E. Kotzé
Next word prediction task is the application of a language model in natural language generation that deals with generating words by repeatedly sampling the next word conditioned on the previous choices. This paper proposes a stacked language model for optimized next word generation using three models. In stage I, the meaning of a word is captured through learn embedding and the structure of the text sequence is encoded using a stacked Long Short Term Memory (LSTM). In stage II, a Bidirectional Long Short Term Memory (Bi-LSTM) stacking on top of the unidirectional LSTM encodes the structure of the text sequences, while in stage III, a two-layer Gated Recurrent Unit (GRU) is used to capture text sequences of data. The proposed system was implemented using Python 3.7, Tensorflow 2.6.0 with Keras and a Nvidia Graphical Processing Unit (GPU). The proposed deep learning models were trained using the Pride and Prejudice corpus from the Project Gutenberg library of ebooks. The evaluation was performed by predicting the next 3 words after considering 10 sets of text sequences. From the experiment carried out, the accuracy of the two-layer LSTM model measured 83%, the accuracy of the Bi-LSTM stacking on unidirectional LSTM model measured 79%, and the accuracy of the two-layer GRU model measured 81%. Regarding predictions, the two-layer LSTM predicted the 10 sequences correctly, the Bi-LSTM stacking on unidirectional LSTM predicted 8 sequences correctly and the two-layer GRU predicted 7 sequences correctly.
{"title":"Stacked Language Models for an Optimized Next Word Generation","authors":"E. O. Aliyu, E. Kotzé","doi":"10.23919/IST-Africa56635.2022.9845545","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845545","url":null,"abstract":"Next word prediction task is the application of a language model in natural language generation that deals with generating words by repeatedly sampling the next word conditioned on the previous choices. This paper proposes a stacked language model for optimized next word generation using three models. In stage I, the meaning of a word is captured through learn embedding and the structure of the text sequence is encoded using a stacked Long Short Term Memory (LSTM). In stage II, a Bidirectional Long Short Term Memory (Bi-LSTM) stacking on top of the unidirectional LSTM encodes the structure of the text sequences, while in stage III, a two-layer Gated Recurrent Unit (GRU) is used to capture text sequences of data. The proposed system was implemented using Python 3.7, Tensorflow 2.6.0 with Keras and a Nvidia Graphical Processing Unit (GPU). The proposed deep learning models were trained using the Pride and Prejudice corpus from the Project Gutenberg library of ebooks. The evaluation was performed by predicting the next 3 words after considering 10 sets of text sequences. From the experiment carried out, the accuracy of the two-layer LSTM model measured 83%, the accuracy of the Bi-LSTM stacking on unidirectional LSTM model measured 79%, and the accuracy of the two-layer GRU model measured 81%. Regarding predictions, the two-layer LSTM predicted the 10 sequences correctly, the Bi-LSTM stacking on unidirectional LSTM predicted 8 sequences correctly and the two-layer GRU predicted 7 sequences correctly.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124183782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/IST-Africa56635.2022.9845650
Dillys Larbi, K. Anthun, F. Asah, O. Debrah, Konstantinos Antypas
The use of electronic health systems is rapidly spreading in low-and middle-income countries (LLMICs). Empirical evidence shows that eHealth systems can improve access, quality, and equitable healthcare delivery, especially for the poor and vulnerable. Studies suggest that a lack of systems thinking leads to inadequate technical infrastructure, lack of interoperability, streamlining of patient-and health information sharing. This article assesses the BETTEReHEALTH strategic priority factors from four African countries: Ethiopia, Ghana, Malawi, and Tunisia. The primary data source was eHealth policies from the four countries. A document analysis was conducted, complemented by deductive, qualitative content analysis. The results show these countries have adopted and implemented eHealth policies. They have dedicated governing bodies that aim to strengthen the coordination of eHealth efforts. However, there is a need for more robust government support and regulation in creating a sustainable national eHealth environment.
{"title":"Assessing Strategic Priority Factors in eHealth Policies of Four African Countries","authors":"Dillys Larbi, K. Anthun, F. Asah, O. Debrah, Konstantinos Antypas","doi":"10.23919/IST-Africa56635.2022.9845650","DOIUrl":"https://doi.org/10.23919/IST-Africa56635.2022.9845650","url":null,"abstract":"The use of electronic health systems is rapidly spreading in low-and middle-income countries (LLMICs). Empirical evidence shows that eHealth systems can improve access, quality, and equitable healthcare delivery, especially for the poor and vulnerable. Studies suggest that a lack of systems thinking leads to inadequate technical infrastructure, lack of interoperability, streamlining of patient-and health information sharing. This article assesses the BETTEReHEALTH strategic priority factors from four African countries: Ethiopia, Ghana, Malawi, and Tunisia. The primary data source was eHealth policies from the four countries. A document analysis was conducted, complemented by deductive, qualitative content analysis. The results show these countries have adopted and implemented eHealth policies. They have dedicated governing bodies that aim to strengthen the coordination of eHealth efforts. However, there is a need for more robust government support and regulation in creating a sustainable national eHealth environment.","PeriodicalId":142887,"journal":{"name":"2022 IST-Africa Conference (IST-Africa)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116893753","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：