University Publishes & Academic Resources
by: Chung Lee, PhD.
Distinguished University Professor, School of Technology Western Covenant University
Title: Emotion Detection from Written Text using Hidden Markov Model
Cite: Lee, Chung. (2022.) Emotion Detection from Written Text using Hidden Markov Model. Mongolian Society of Science and Technology. pp. 211-216
DOI: https://www.academia.edu/104773312/HMM_for_Emotion_Detection_Lee
Abstract:
This research focuses on integrating emotion into Human-Computer Interaction (HCI) and chatbot implementations through the detection of emotions from written texts. Emotion recognition is challenging in text-based communication as it lacks nonverbal cues. Various methodologies, including keyword-based and statistical NLP-based detection, have been explored but with limited success. The study proposes a new approach using the Hidden Markov Model (HMM) for emotion detection, where written texts are manually tagged into predefined emotion categories such as angry, happy, sad, and neutral.
The researchers developed the Emotion Detection Tool (EDT), a Java-based software that accepts natural language sentences and outputs the detected emotion. They conducted verification tests using two types of samples: sentences already in the database and new sentences with known emotions. The results demonstrated that the POS-based detection outperformed word-based detection, highlighting the validity of the HMM approach in emotion detection from text. The EDT surpassed keyword-based detection, allowing detection from subsets of sentences, such as clauses, phrases, or words. The researchers propose further improvements by incorporating big data and deep learning techniques, as they expect these modern AI trends to yield more accurate emotion detection results.
by: Junho Oh, PhD.
Professor, School of Technology Western Covenant University
Title: A scalable method for detecting multiple loci associated with traits using TF-IDF weighting and association rule mining
Cite: Sunwon Lee, Jaewoo Kang and Junho Oh. A Scalable Method for Detecting Multiple Loci Associated with Traits using TF-IDF Weighting and Association Rule Mining. IEEE International conference on Bioinformatics and Biomedicine Workshops (BIBMW). Hong Kong, China. Dec 18-21 2010. pp. 318-323.
DOI: https://doi.org/10.1109/BIBMW.2010.5703821
Abstract:
The recent advance in SNP genotyping has made a significant contribution to reduction of the costs for large-scale genotyping. The development also has dramatically increased the size of the SNP genotype data. The increase of the volume of the data, however, has posed a huge obstacle to the conventional analysis techniques that are typically vulnerable to the high-dimensionality problem. To address the issue, we propose a method that exploits two well-tested models: the document-term model and the transaction analysis model. The proposed method consists of two phases. In the first phase, we reduce the dimensions of the SNP genotype data by extracting significant SNPs through transformation of the data in lieu of the document-term model. In the second phase, we discover the association rules that signify the relations between the SNPs and the traits, through the application of the transactional analysis in the reduced-dimension genotype data. We validated the discovered rules through the literature survey. Experiments were also carried out using the HGDP panel data provided by the Foundation Jean Dausset-CEPH, which prove the validity of our new method for identifying appropriate dimensional reduction and associations of multiple SNPs and traits.
by: Junho Oh, PhD.
Professor, School of Technology Western Covenant University
Title: A 2-phased approach for detecting multiple loci associations with traits
Cite: Sunwon Lee, Jaewoo Kang, & Junho Oh. (2012). A 2-phased approach for detecting multiple loci associations with traits. International Journal of Data Mining and Bioinformatics. Vol. 6. Issue 5. Pages 535-556.
DOI: https://doi.org/10.1504/IJDMB.2012.049318
Abstract:
The recent advance in SNP genotyping has made a significant contribution to reduction of the costs for large-scale genotyping. The development also has dramatically increased the size of the SNP genotype data. The increase in the volume of the data, however, has posed a huge obstacle to the conventional analysis techniques that are typically vulnerable to the high-dimensionality problem. To address the issue, we propose a method that exploits two well-tested models: the document-term model and the transaction analysis model. The proposed method consists of two phases. In the first phase, we reduce the dimensions of the SNP genotype data by extracting significant SNPs through transformation of the data in lieu of the document-term model. In the second phase, we discover the association rules that signify the relations between the SNPs and the traits, through the application of transactional analysis in the reduced-dimension genotype data. We validated the discovered rules through literature survey. Experiments were also carried out using the HGDP panel data provided by the Foundation Jean Dausset-CEPH, which prove the validity of our new method for identifying appropriate dimensional reduction and associations of multiple SNPs and traits. This paper is an extended version of our workshop paper presented in the 2010 International Workshop on Data Mining for High Throughput Data from Genome-Wide Association Studies.
by: Benjamin Olufemi Adjumo
MAR Candidate, School of Religious Studies. Western Covenant University
Title: Modelling Drug Addictions and Associated Risk Factors in California
Cite: Adjumo, B. (2023). Modelling the drug addictions and associated risk factors in California. International Journal of Pharmacology and Clinical Research. Vol. 5(1): pp. 17-21.
DOI: https://doi.org/10.33545/26647613.2023.v5.i1a.20
Abstract:
Drug addiction is a chronic medical condition marked by an individual’s inability to suppress the need to use drugs despite knowing that doing so will have a severe impact on their health, their family, and society at large. A growing corpus of research demonstrates how the use of illegal drugs begins and persists as a result of interactions between the substance misuse phenomena and drug addiction propensity. The purpose of this study was to evaluate the relationship between socio demographic factors and other risk factors and the propensity for drug addiction. According to the results of the fitted logistic regression, there is a significant correlation between the likelihood of drug addiction in California and social demographic risk factors like sample respondents, family history, alcohol consumption, monthly income, age, and gender as well as other risk factors like tobacco smoking. The goodness of fit test shows that the model fits the data well and may be used to forecast the likelihood of drug addiction in the future. Based on the fact that California residents of all ages particularly adults and genders are a significant factor in the occurrence of drug addiction, the state should enforce rules and consistently raise public knowledge of potential penalties and fines.
Keywords: Drug addictions, Socio-demographic risk factors, logistic regression model, goodness of fit
by: Benjamin Olufemi Adjumo
MAR Candidate, School of Religious Studies. Western Covenant University
Title: Effective Corporate Communications as a Catalyst for Organization Profitability, Growth and Sustainability.
Cite: Adjumo, B. and Jaiyeola, O. (2023). Effective Corporate Communications as a Catalyst for Organization Profitability, Growth and Sustainability. European Journal of Business and Management. Vol 15, No. 6. Pp. 48-55.
DOI: https://doi.org/10.7176/EJBM/15-6-05
Abstract:
Every organization exists for the purpose of actualizing the set-out goals and objectives as particularly enshrined in expansion and sustainability. One of the ways all these goals and objectives of the organization can translate to visible results is through effective corporate communication which clearly outlines the very purpose of the organization and the language of the organization. What corporate communication does is the linking together of every aspect of the organization in the pursuit of the goals of the organization by establishing contacts with all those involved. Corporate communication ensures that both the internal and external stakeholders in an organization are effectively carried along and that is the very key reason the concept is considered as highly imperative towards the growth of the organization. Its ability to analyze environment, come up with policy and actions, ensure the spirit of teamwork, protect and improve the corporate identity of the organization amongst others, clearly underscore the importance and uniqueness of corporate communication as not just the machinery that can ensure organization’s growth but also a tool that can help sustain the growth of an organization.
Keywords: Sustainability Communications Catalyst Profitability Growth Sustainability