Patricia E. Nalwoga Lutu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Patricia E. Nalwoga Lutu has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- Andries Petrus Engelbrecht
Department of Industrial Engineering and Computer Science Division, Stellenbosch University, Stellenbosch 7600, South Africa
3 shared publications - Refiloe Motjolopane
School of Information Technology University of Pretoria Pretoria South Africa
1 shared publication V.V. Nhlabano
Department of Computer Science, University of Pretoria, Pretoria, South Africa
1 shared publication
Latest Publications
- 1 November 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 155-159
Using Twitter Mentions and a Graph Database to Analyse Social Network Centrality
Social networks are one category of social media that facilitates the formation of communities, sharing of content, and meeting people. Twitter is a popular microblogging and social networking service. Social media marketers within business organisations, are interested in identifying popular social network users, known as influencers who can be targeted for purposes of word-of-mouth branding. For Twitter, influencers are those users who have many followers. Influencers are typically identified through graph mining of social networks data. This type of mining involves the analysis of links between the graph nodes which store data for social network members. Follows relationships are commonly used to analyse Twitter social networks. The purpose of this paper is to demonstrate how mentioned relationships in Twitter data can be used to create a social networks graph database. Centrality measures are then used to analyse the social networks. It is demonstrated that the analysis of social networks based on the mentioned relationships can provide more information about influencers compared to the analysis of social networks based on the follows relationships.
- 1 November 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 155-159
Using Twitter Mentions and a Graph Database to Analyse Social Network Centrality
Social networks are one category of social media that facilitates the formation of communities, sharing of content, and meeting people. Twitter is a popular microblogging and social networking service. Social media marketers within business organisations, are interested in identifying popular social network users, known as influencers who can be targeted for purposes of word-of-mouth branding. For Twitter, influencers are those users who have many followers. Influencers are typically identified through graph mining of social networks data. This type of mining involves the analysis of links between the graph nodes which store data for social network members. Follows relationships are commonly used to analyse Twitter social networks. The purpose of this paper is to demonstrate how mentioned relationships in Twitter data can be used to create a social networks graph database. Centrality measures are then used to analyse the social networks. It is demonstrated that the analysis of social networks based on the mentioned relationships can provide more information about influencers compared to the analysis of social networks based on the follows relationships.
- 1 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Improving Classification Performance for Minority Classes through the use of Positive-Versus-Negative Ensemble Models
Classification is a commonly used modelling method for data mining. A classification model is a predictive model which is used to predict a categorical value, called a class. Ensemble classification modelling involves the creation of several base models and a combination algorithm for the base model predictions. The classification modelling process uses a set of instances called the training data. Each instance consists of values for the predictor variables and a categorical label called the class. A class is called a minority class if it has a much smaller number of training instances compared to the other classes. This results in a low level of correct classification compared to the classification performance for the majority classes. Positive-versus-negative (pVn) classification has been reported in the literature as an ensemble classification method which is applicable to classification modelling for multi-class prediction tasks. The purpose of this paper is to report on experimental results for the performance of a replication method for improving classification performance for minority classes, using pVn classification modelling. The experimental results demonstrate that the classification accuracy for minority classes can be improved through the use of pVn classification models.
- 1 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Improving Classification Performance for Minority Classes through the use of Positive-Versus-Negative Ensemble Models
Classification is a commonly used modelling method for data mining. A classification model is a predictive model which is used to predict a categorical value, called a class. Ensemble classification modelling involves the creation of several base models and a combination algorithm for the base model predictions. The classification modelling process uses a set of instances called the training data. Each instance consists of values for the predictor variables and a categorical label called the class. A class is called a minority class if it has a much smaller number of training instances compared to the other classes. This results in a low level of correct classification compared to the classification performance for the majority classes. Positive-versus-negative (pVn) classification has been reported in the literature as an ensemble classification method which is applicable to classification modelling for multi-class prediction tasks. The purpose of this paper is to report on experimental results for the performance of a replication method for improving classification performance for minority classes, using pVn classification modelling. The experimental results demonstrate that the classification accuracy for minority classes can be improved through the use of pVn classification models.
- 1 August 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Impact of Text Pre-Processing on the Performance of Sentiment Analysis Models for Social Media Data
Sentiment analysis also known as opinion mining is the study of people's opinions, attitudes, appraisals and emotions regarding events, individuals, issues, entities, topics and their attributes. The concept of Sentiment Analysis can also be extended to include detecting emotional status such as sadness, anger, and happiness. The purpose of Sentiment Analysis is to predict the polarity of a given text item in order to determine if the author is expressing positive, negative or a neutral opinion about a given topic. Sentiment Analysis has become an interesting research topic over the years due to a wide variety of its practical applications, its promising commercial benefits and also due to the many interesting challenges and research problems it presents to the research field. In Sentiment analysis one of the most crucial step is text pre-processing which is treated as a challenging text classification task as it classifies the orientation of a text into either positive or negative. In this study we investigates the influence of Text pre-processing methods used for Social media data. The experimental results demonstrate that Text pre-processing methods increases the predictive accuracy of the resulting models for sentiment classification.
- 1 August 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Impact of Text Pre-Processing on the Performance of Sentiment Analysis Models for Social Media Data
Sentiment analysis also known as opinion mining is the study of people's opinions, attitudes, appraisals and emotions regarding events, individuals, issues, entities, topics and their attributes. The concept of Sentiment Analysis can also be extended to include detecting emotional status such as sadness, anger, and happiness. The purpose of Sentiment Analysis is to predict the polarity of a given text item in order to determine if the author is expressing positive, negative or a neutral opinion about a given topic. Sentiment Analysis has become an interesting research topic over the years due to a wide variety of its practical applications, its promising commercial benefits and also due to the many interesting challenges and research problems it presents to the research field. In Sentiment analysis one of the most crucial step is text pre-processing which is treated as a challenging text classification task as it classifies the orientation of a text into either positive or negative. In this study we investigates the influence of Text pre-processing methods used for Social media data. The experimental results demonstrate that Text pre-processing methods increases the predictive accuracy of the resulting models for sentiment classification.
- 1 August 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
The use of Histogram Analysis to Support Fast Selection of Predictive Features for Data Stream Mining
A data stream is a real-time, continuous, ordered sequence of items. Many stream mining methods for fast data selection and modeling from large quantities of stream data have been proposed in the literature. Histograms are commonly used in various data stream mining activities such as exploratory data analysis and pre-processing of training data for classification algorithms. This paper reports on an exploratory study of the use of histograms as a basis for fast selection of relevant features for predictive classification. Histograms are analysed using simple statistical measures in order to establish feature relevance. Experimental results demonstrate that the proposed method has a high potential of selecting relevant features which result in high predictive performance.
- 1 August 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
The use of Histogram Analysis to Support Fast Selection of Predictive Features for Data Stream Mining
A data stream is a real-time, continuous, ordered sequence of items. Many stream mining methods for fast data selection and modeling from large quantities of stream data have been proposed in the literature. Histograms are commonly used in various data stream mining activities such as exploratory data analysis and pre-processing of training data for classification algorithms. This paper reports on an exploratory study of the use of histograms as a basis for fast selection of relevant features for predictive classification. Histograms are analysed using simple statistical measures in order to establish feature relevance. Experimental results demonstrate that the proposed method has a high potential of selecting relevant features which result in high predictive performance.
- 1 December 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 335-340
Naïve Bayes Classification Ensembles to Support Modeling Decisions in Data Stream Mining
Data stream mining is the process of applying data mining methods to a data stream in real-time in order to create descriptive or predictive models. Due to the dynamic nature of data streams, new classes may emerge as a data stream evolves, and the concept being modelled may change with time. This gives rise to the need to continuously make revisions to the predictive model. Revising the predictive model requires that labelled training data should be available. Manual labelling of training data may not be able to cope with the speed at which data needs to be labelled. This paper proposes a predictive modeling framework which supports two of the common decisions that need to be made in stream mining. The framework consists of two components: an online component and an offline component. The online component uses Naïve Bayes ensemble base models to make predictions for newly arrived data stream instances. The offline component consists of algorithms to combine base model predictions, determine the reliability of the ensemble predictions, select training data for new base models, create new base models, and determine whether the current online base models need to be replaced.
- 1 December 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 335-340
Naïve Bayes Classification Ensembles to Support Modeling Decisions in Data Stream Mining
Data stream mining is the process of applying data mining methods to a data stream in real-time in order to create descriptive or predictive models. Due to the dynamic nature of data streams, new classes may emerge as a data stream evolves, and the concept being modelled may change with time. This gives rise to the need to continuously make revisions to the predictive model. Revising the predictive model requires that labelled training data should be available. Manual labelling of training data may not be able to cope with the speed at which data needs to be labelled. This paper proposes a predictive modeling framework which supports two of the common decisions that need to be made in stream mining. The framework consists of two components: an online component and an offline component. The online component uses Naïve Bayes ensemble base models to make predictions for newly arrived data stream instances. The offline component consists of algorithms to combine base model predictions, determine the reliability of the ensemble predictions, select training data for new base models, create new base models, and determine whether the current online base models need to be replaced.
- 2 September 2014
- Published by Springer Nature
- p. 103-124
Web 2.0 Computing and Social Media as Solution Enablers for Economic Development in Africa
- 2 September 2014
- Published by Springer Nature
- p. 103-124
Web 2.0 Computing and Social Media as Solution Enablers for Economic Development in Africa
- 1 March 2012
- Published by Elsevier in Expert Systems with Applications
- Vol. 39 (4) , 4358-4376
Using OVA modeling to improve classification performance for large datasets
- 1 March 2012
- Published by Elsevier in Expert Systems with Applications
- Vol. 39 (4) , 4358-4376
Using OVA modeling to improve classification performance for large datasets
- 1 January 2012
- Published by Wiley in The Electronic Journal of Information Systems in Developing Countries
- Vol. 50 (1) , 1-22
Improving Data Quality in the Banking Supervisory Data of Southern Africa Central Banks
The importance of effective supervision of banking institutions in developing countries cannot be over‐emphasised. Central banks are responsible for the supervision of the banking institutions of a country. In order to effectively perform the supervisory role of banking institutions, central banks need to collect and analyse data about the operations of the banking institutions they supervise. Data analysis is conducted in order to establish the health of these institutions. Central banks in the SADC region experience many problems with the quality of the data that they receive from the banking institutions they supervise. Needless to say, decisions made on the basis of poor quality data increase the risk of allowing unsound institutions to continue in operation. This creates the real risk of loss of money by the individuals and organisations that form the clientele of these institutions, and will adversely affect the economy of the country. Despite the commendable efforts made by the regional central banks to implement a banking supervision application to expedite the supervisory activities through instituting a one‐stop‐shop for all the banking supervisory information, and the introduction of an electronic platform for collecting and analysing periodic banking supervision returns, there are still many imperfections in the information. A survey was conducted to investigate data quality problems at three SADC central banks. Recommendations are provided in this paper on measures that can be taken to improve the quality of the banking supervisory data for the SADC central banks.
- 1 January 2012
- Published by Wiley in The Electronic Journal of Information Systems in Developing Countries
- Vol. 50 (1) , 1-22
Improving Data Quality in the Banking Supervisory Data of Southern Africa Central Banks
The importance of effective supervision of banking institutions in developing countries cannot be over‐emphasised. Central banks are responsible for the supervision of the banking institutions of a country. In order to effectively perform the supervisory role of banking institutions, central banks need to collect and analyse data about the operations of the banking institutions they supervise. Data analysis is conducted in order to establish the health of these institutions. Central banks in the SADC region experience many problems with the quality of the data that they receive from the banking institutions they supervise. Needless to say, decisions made on the basis of poor quality data increase the risk of allowing unsound institutions to continue in operation. This creates the real risk of loss of money by the individuals and organisations that form the clientele of these institutions, and will adversely affect the economy of the country. Despite the commendable efforts made by the regional central banks to implement a banking supervision application to expedite the supervisory activities through instituting a one‐stop‐shop for all the banking supervisory information, and the introduction of an electronic platform for collecting and analysing periodic banking supervision returns, there are still many imperfections in the information. A survey was conducted to investigate data quality problems at three SADC central banks. Recommendations are provided in this paper on measures that can be taken to improve the quality of the banking supervisory data for the SADC central banks.