Worarat Krathu

Semi-automatic construction of thyroid cancer intervention corpus from biomedical abstracts

Thyroid cancer is a common endocrine tumor that is experiencing a steady increase in incidence worldwide. The latest discoveries on disease and its treatment are mostly propagated in the form of biomedical publications such as those in PubMed. Unfortunately, this information is distributed in unstructured text with over two thousand articles being added annually. Text mining technology plays an important role in information extraction, since it can be used to uncover hidden value from the vast amount of text in reasonable time. In general, a preliminary task of text mining is Named Entity Recognition (NER). In this case, a gold standard corpus is needed, since the capability of NER depends on a trustworthy corpus. However the construction of gold standard corpus is a laborious and time-consuming process. In order to obtain a reasonably practical corpus in a limited time, this paper consequently proposes a semiautomatic approach to construct a thyroid cancer interventions corpus. The experimental results demonstrate that the proposed method can be used to construct a thyroid cancer intervention corpus reasonably in terms of both performance and overfitting avoidance.

Data mining approach for automatic discovering success factors relationship statements in full text articles

In the context of Business-to-Business (B2B), an understanding of inter-organizational success factors and their impacts is crucial for effective strategic management. Several studies regarding those success factors and their influences have been conducted and published as articles. We aim at applying existing techniques, especially data mining, to automatically classify relevant sentences describing an influencing relationship between success factors. This paper presents the experiment method and results to find the optimal data mining workflow for our classification task. In particular, we apply several well-known data mining techniques based on different control factors. Then all discovered models are evaluated and compared to find the optimal data mining workflow. The main contributions include (i) the application of data mining for discovering success factors and their relationships, and (ii) the optimal workflow as a standardized flow for further similar classification tasks. The major challenge of this work is that there exists no mature corpus in this context, and hence our approach is implemented without a supporting corpus. The result shows that the models derived from the workflows that consider a section where a sentence is located perform better than the others in term of average performance. Furthermore, we found that the Support Vector Machine (SVM) performs better than other classifiers.

Network analysis of inter-organizational success factor relationships

One of the best ways to deal with the problem of knowledge distillation in unstructured text is applying text mining. This machine learning-based approach can provide extracted useful information from large body of texts, in a reasonable time. However, the results are usually in complicated forms, which meant it is a non-trivial task in term of interpretation. In this paper, we present appropriate visualizations and analyses in order to tackle the tangled network representing relationship between entities (i.e. terms extracted from raw text). Conducted on a case study of information extraction in the business management domain, our results reveal the hidden relationships of inter-organizational success factors in a simple structured way. These results can assist companies to understand and generate business strategies, in terms of the collaboration aspect.

Hadoop as a Service for Academic Purposes

The growth of data science education at both graduate and undergraduate levels has increase dramatically. Need for computing platforms that are well-equip with the appropriate tools for students and instructors, and that are readily accessible and flexibly configurable has subsequently become more conspicuous. Using lots of sandboxes that would consume tremendous resources is not an effective solution. This paper propose a solution that design and implement Hadoop as a private online service that could effectively meet the design characteristics of an academic data science computing platform. The propose design emphasise a quick deployment solution based on available resources and technologies. The case study that follow demonstrate and cultivate our approach as a proof of concept for any academic institutions to implement their own services. This paper contains many aspects of service implementation from managing resources to choosing the right Hadoop platform components and solution to accounting for user authentication and authorisation as well as security and network design.

Automatic discovering success factor relationship entities in articles using named entity recognition

An understanding of success factor relationships in the context of business-to-business where Inter-organizational Relationship (IORs) between organizations is crucial for effective strategic management to accomplish marketing goals. Several studies regarding those success factors and their influences have been conducted and published as articles. We apply the technique of Named Entity Recognition and find a suitable model for extracting the entities of success factors. The appropriate model needs only 60 research abstracts and the performance as high as 0.9 is achievable. Furthermore, we find that there is no significant improvement affected by applying sentence normalization.

Meerkat: A framework for developing presence monitoring software based on face recognition

Presence monitoring is the feature that recognizes presence of a person or persons automatically. It improves upon both manual and automatic presence verification process by allowing unobtrusive and continuous monitoring rather than performing a discrete check at the beginning and/or at the end of the participation period. The software with presence monitoring capability is particularly useful in todays higher education settings where various soft skills must be continuously developed, monitored and assessed. This paper proposes a framework-called Meerkat-for developing presence monitoring software. The framework relies on the face recognition technology from Microsoft Cognitive Services as a convenient tool to produce web-based APIs that can easily be used to develop web applications for presence monitoring. In a case study, an application has been developed as a proof of concept to confirm the integration between the presence monitoring feature and the Face API of Microsoft Cognitive Services. Furthermore, to evaluate the performance of the application, an error analysis on this application has been carried out that shows a satisfactory performance. As Meerkat is based on face recognition which extends the Microsoft Cognitive Services, the results confirm that most of the errors highly correlate with the image quality and the posture of the faces.

Metabolite Named Entity Recognition: A Hybrid Approach

Since labor intensive and time consuming issue, manual curation in metabolic information extraction currently was replaced by text mining (TM). While TM in metabolic domain has been attempted previously, it is still challenging due to variety of specific terms and their meanings in different contexts. Named Entity Recognition (NER) generally used to identify interested keyword (protein and metabolite terms) in sentence, this preliminary task therefore highly influences the performance of metabolic TM framework. Conditional Random Fields (CRFs) NER has been actively used during a last decade, because it explicitly outperforms other approaches. However, an efficient CRFs-based NER depends purely on a quality of corpus which is a nontrivial task to produce. This paper introduced a hybrid solution which combines CRFs-based NER, dictionary usage, and complementary modules (constructed from existing corpus) in order to improve the performance of metabolic NER and another similar domain.

Telemedicine acceptance framework for the elderly in Thailand

Thailand is currently moving toward an ageing society. According to the current population structure, the number of elderly tends to be increasing while the number of working-age people is decreasing. Often the elderly require intensive healthcare, and due to the growing elderly population, the need for an adequate number of healthcare professionals to provide healthcare services becomes a big issue. Telemedicine has emerged as a solution by providing professional healthcare services in remote areas. However, telemedicine in Thailand is still underutilized for many reasons. In this work, we studied factors that affect the acceptance of telemedicine for the elderly in Thailand. In particular, we applied the UTAUT framework for confirming the relationships between the proposed factors and the behavioral intention. The results provide an insight into understanding how the elderly accept telemedicine, and hence, enable the design of the telemedicine acceptance framework for the elderly in Thailand.