Soyeon Caren Han

Using MCRDR based Agile approach for expert system development

Various expert system development approaches were proposed but most of them cannot deal with two problems: the difficulty of analysis and maintenance. Rather than to spend time waiting any longer, it is better to find an alternative solution from other research fields. In computer software development area, researchers have been suffering from the difficulty of maintenance and analysis, just as the researchers in the expert system development field. To solve this issue, researchers in the software used both agile software development and business rules approach: agile software development is for overcoming the the difficulty of analysis, and business rules approach is for reducing issues in the maintenance. There is a big opportunity that those two approaches can also be solve the two issues in the expert system development field. The paper describes requirements of the approach based on agile software development and the business rules approach. As a result, we consider and specify why the Multiple Classification Ripple Down Rules is the novel approach for the expert system development.

Social issue gives you an opportunity: discovering the personalised relevance of social issues

Social networking services have received a lot of attention recently so that the discussion of certain issues is becoming more dynamic. Many websites provide a new service that displays the list of the trending social issues. It is very important to respond to those social issues since the impact on organisations or people may be considerable. In this paper, we present our research on developing the personalised relevance identification system that displays the relevance of social issues to a target domain. To accomplish this, we first collected social issue keywords from Google Trends, Twitter and Google News. After that, we setup an electronic document management system as a target domain that would include all knowledge and activities having to do with a target object. In order to identify the relevance of the social issues to a target, we applied the Term Frequency Inverse Document Frequency (TFIDF). Our experiments prove that we can identify the meaningful relevance of social issues to targets, such as individuals or organizations.

Identifying the Relevance of Social Issues to a Target

Responding to social issues is very crucial because their impact can be significant to organizations or individuals. In this paper, we focus on proposing the method that identifies the personalized relevance of social issues to targets, such as individuals or organizations. To achieve this aim, we first collected trending social issues from Google Trends, microblog, and Internet news. Then, we obtained the well-structured document management system as a target domain that contains all activities regarding target objects. We applied the Term Frequency Inverse Document Frequency to obtain the personalized relevance weight of the social issue to a target.

It Is Time to Prepare for the Future: Forecasting Social Trends

A social issue is what arises when the public discuss a specific event. Recently, many large Internet based service companies provide new trends services that display the emerging issues based on their data, for example, Google displays “top 10 most searched topics” every hour. Those emerging issues reflect the trend of public interest. Forecasting those issues helps the user to prepare for the future. In this paper, we present our research on proposing the social issue-forecasting model. To do so, we first collected social issue keyword from Google Trends for 3 months since it is based on the large scale of public data. We apply the k-nearest neighbor algorithm, which is the pattern recognition technology for recognizing the complex patterns and trends. To improve the accuracy, we applied Ripple Down Rules.

Exploring a role for MCRDR in enhancing telehealth diagnostics

In-home telehealth devices are becoming increasingly popular when it comes to supporting the health management of home-based patients. With the devices capable of highly active monitoring, using sensors which produce large amounts of data, the deployment of telehealth devices into the home highlights the need for improved ways to collate, classify and dynamically interpret data safely and effectively. For clinicians working at a distance, the amounts of data generated by all in-home patient telematics devices poses questions on how best to intelligently filter, analyze and interpret this data to make diagnoses and respond to changes in patient conditions. In order to manage this issue, expert systems, applied for decades in other health fields, might play a role. In this paper, we explore how one type of expert system, Multiple Classification Ripple Down Rules (MCRDR), might address the issues. This paper begins by reviewing the capabilities of expert systems. Specifically, MCRDR is reviewed and its integration with an example telehealth device, MediStation, is explored. The range of potential benefits which might accrue when MCRDR and the MediStation are linked is identified as are some research and development challenges. Moving forwards, a simple simulator is introduced as one approach which is shown to be effective at exploring this exciting area of research. This paper takes the first steps towards introducing expert systems into the uHealth field and presents a simulator for this purpose.

Twitter Trending Topics Meaning Disambiguation

Twitter is one of the most popular social media services that allow users to share and spread information. Twitter monitors their users’ postings and detects the most discussed topics of the moment. Then, they publish these topics on the list, called ‘Trending Topics’. Trending Topics on Twitter shows the list of top 10 trending topics but each topic consists of short phrase or keyword, which does not contain any explanation of those meanings. It is almost impossible to identify what a trending topic is about unless you read all related tweets. The goal of this paper is finding the most successful method that uses to retrieve the representative contents of trending topics in order to disambiguate the meaning of topics. We first collected the trending topics and tweets related to them. Then, we applied four types of information retrieval approaches (key factor extraction, named entity recognition, topic modelling, and automatic summarization) for extracting the representative contents of trending topics. We conducted human experiments with 20 postgraduate students.

Trending Topics Rank Prediction

Many web services, such as Twitter and Google, provide a list of their most popular terms, called a trending topics list, in descending order of popularity ranking. The changes in people’s interest in a specific trending topic are reflected in the changes of its popularity rank (up, down, and unchanged). This paper analyses the nature of trending topics and proposes a temporal modelling framework for predicting rank change of trending topics using historical rank data. Historical rank data show that almost 70 % of trending topics tend to disappear and reappear later. Therefore it is important to reflect this phenomenon in the prediction model, which is related to handling missing value and window size. Missing value handling approach was selected by using expectation maximization. An optimal window size is selected based on the minimum length of topic disappearance in the same topic but with a different context. We examined our approach with four machine-learning techniques using the U.S. twitter trending topics collected from 30th June 2012 to 30th June 2014. Our model achieved the highest prediction accuracy (94.01 %) with C4.5 decision tree algorithm.

V&V to Use Agile Approach in ES Development: Why RDR Works for Expert System Developments!

In artificial intelligence, many researchers have proposed several expert system development approaches but most of them failed to deal with two issues, maintenance and analysis. It is better to find an alternative solution from other areas, rather than to waste time waiting any longer. We found that researchers in computer software development also have been suffering from the difficulty of maintenance and analysis, just as in the expert system development area. To solve this problem, agile software development is used to overcome the difficulty of analysis, and business rules approach is utilised for removing maintenance issues. We believe that the two approaches are the ideal solutions that are able to formalize the expert system development process. In this paper, we outline this novel approach, Multiple Classification Ripple Down Rule, which is based on agile software development and business rules approach.

Machine learning for the peer assessment credibility

The peer assessment approach is considered to be one of the best solutions for scaling both assessment and peer learning to global classrooms, such as MOOCs. However, some academic staff hesitate to use a peer assessment approach for their classes due to concerns about its credibility and reliability. The focus of our research is to detect the credibility level of each assessment performed by students during peer assessment. We found three major scopes in assessing the credibility level of evaluations, 1) Informativity, 2) Accuracy, and 3) Consistency. We collect assessments, including comments and grades provided by students during the peer assessment process and then each feedback-and-grade pair is labeled with its credibility level by Mechanical Turk evaluators. We extract relevant features from each labeled assessment and use them to build a classifier that attempts to automatically assess its level of credibility in C5.0 Decision Tree classifier. The evaluation results show that the model can be used to automatically classify peer assessments as credible or non-credible, with accuracy in the range of 88%.

RDR-based knowledge based system to the failure detection in industrial cyber physical systems

Cyber Physical System(CPS) allows to collect different sensor and alarm data from large number of facilities in industrial plants. Failure and faulty diagnosis is one of the most complicated and dynamic problems in the industrial plant management since most of failures are extremely ambiguous which needs to be solved based on an expert’s experience. This makes the solutions very subjective and requires too much time, efforts and monetary investment. In this paper, we are proposing new failure detection approach with machine learning and human expertise by using alarm data. As the first step of developing this new method, we collected several types of alarm data that detected functional failure in Hyundai Steel factory. We analyzed and processed the alarm data with 35 domain experts. Based on the data, we propose a knowledge based system which is Ripple Down Rule-based. This system acquires knowledge by machine learning which is maintained by human experts. The evaluation results showed that the proposed failure detection framework can reduce the time of human expertise acquisition and the cost of solving over-generalization and over-fitting problems by using machine learning techniques.