Chuan-Chun Wu

Targeting customers via discovery knowledge for the insurance industry

In this paper, the knowledge discovery in databases and data mining (KDD/DM), one of the data-based decision support technologies, is applied to help in targeting customers for the insurance industry. In most KDD/DM application cases, major tasks are required, including data preparation, data preprocessing, data mining, interpretation, application and evaluation. A case study is presented that KDD/DM is utilized to explore decision rules for a leading insurance company. The decision rules can be used to investigate the potential customers for an existing or new insurance product. The research firstly constructed the application framework, then defined and conducted each task required, and finally obtained feedback from the case company. Discussions and implications with respect to this research are presented also.

Examining Retailing Performance Via Financial Index

The modern retailing industry has shown a great impact to the enhancement of economy in Taiwan. In consequence, its financial structure becomes an important issue as long as the performance measurement is concerned. The current research described in this paper contains two parts. The first part is to present an empirical investigation where financial index is considered to measure the retailing performance. The second part is to examine the effects of financial index on sales and gross margin. The technique of Data Envelopment Analysis (DEA) is utilized to derive the efficiency and multiple regression analysis to disclose the effects. The inputs as the independent factors are current assets, the number of employees, inventory investment, and promotion expenses used in the analysis while the outputs as the dependent factors are sales and gross margin. There were remarkable results: (1) the averaged efficiency was 0.7454 which, generally, did not show a good performance, (2) retailers in the group of vehicle parts and supplies performed the best, (3) retailers in the group of fabrics, clothes, and apparel accessories performed the worst, (4) current assets, the number of employees, and promotion expenses showed significant effects on both sales and gross margin. Other research results and findings were addressed and implications discussed in this research also.

Toward intelligent data warehouse mining: An ontology-integrated approach for multi-dimensional association mining

A data warehouse is an important decision support system with cleaned and integrated data for knowledge discovery and data mining systems. In reality, the data warehouse mining system has provided many applicable solutions in industries, yet there are still many problems causing users extra problems in discovering knowledge or even failing to obtain the real and useful knowledge they need. To improve the overall data warehouse mining process, we present an intelligent data warehouse mining approach incorporated with schema ontology, schema constraint ontology, domain ontology and user preference ontology. The structures of these ontologies are illustrated and how they benefit the mining process is also demonstrated by examples utilizing rule mining. Finally, we present a prototype multidimensional association mining system, which with intelligent assistance through the support of the ontologies, can help users build useful data mining models, prevent ineffective pattern generation, discover concept extended rules, and provide an active knowledge re-discovering mechanism.

An Active Multidimensional Association Mining Framework with User Preference Ontology

Business data are subject to change by time or by the modifications of business rules. New knowledge needs to be extracted to reflect the most up to date situations hence periodic or occasional re-mining is essential. This paper proposes an active multidimensional association mining framework that incorporates with user preference ontology, which contains surrogate queries that represent frequently used queries in the query history log. The representative power and the user preference of the surrogate queries are derived and expressed in fuzzy linguistic terms. The construction of the ontology is demonstrated. How it can assist the active mining mechanism is also described. Specifically, the connection of the user preference ontology to the user profile in the enterprise database allows dispatching of new mining results to specific users automatically. A prototype implementation of the proposed system framework is provided and an effectiveness experiment for the user preference ontology is also conducted.

Location-aware service applied to mobile short message advertising

Location-aware recommendation system for mobile short message advertising is proposed.System parameters are collected from storeowners and mobile phone users.Pull-oriented advertising model was used in the prototype system.Implicit content-based mechanism was embedded in the prototype for filtering and comparison.Evaluation on usefulness, mechanism, and privacy revealed acceptable, and feedbacks addressed. The use of location-aware services to deliver short message advertisements requires an in-depth review of the literature in general and the applicable domains and needs and wants of both marketers and consumers in particular. This paper presents a location-aware service application that develops a recommendation system (RS) for mobile short message advertising (LARSMA). The LARSMA prototype is developed through two stages: system parameter consultation and system development. System parameter consultation involves the collection of advertising categories and their attributes from storeowners and current interests and preferences from mobile phone users. Using a pull-oriented advertising model, system development involves the creation of the LARSMA prototype system, which allows storeowners to create and directly send short messages and mobile phone users within the defined area to automatically retrieve short messages that depend on their inputs. The implicit content-based recommendation mechanism embedded in the designed LARSMA performs filtering and comparison to derive the final top N formatted advertising messages as outputs of LARSMA. An application case is used to demonstrate the proposed LARSMA system prototype. Validation and evaluation are conducted, and feedbacks are addressed.

Favorable support threshold recommendation for multidimensional association mining using user preference ontology

The classical algorithms for mining association rule require the user to specify a support threshold to determine if an itemset is frequent or not. Unfortunately, the setting of support threshold is subjective without clear standard and has great influence on the mining results. In this paper we propose an intelligent minimum support suggestion framework with the help of the user preference ontology. The user preference ontology maintains the frequently used mining queries extracted from the mining log. The system finds the most similar queries to the users mining intension, aggregates them and obtains the favorable support range for the user to refer. In this paper we describe briefly the construction of the user preference ontology and focus on the methodology for query similarity comparison.

Ontology-Assisted Query Formulation in Multidimensional Association Rule Mining

In the information era, the development of various electronic information resources have dramatically grown, mining useful information from large databases has become one of the most important issues in information research for users. Information technologies have provided many applicable solutions, yet there are still many problems that cause users to spend extra time to get real knowledge. In this paper, we show an ontology based system framework for multi-dimensional association rule mining that incorporates ontologies in order to help users develop correct queries, reduce the system resource consumption and improve the efficiency of the mining process.

Extracting Knowledge from Open Data of Epidemic Diseases: The Case of Dengue, Influenza, and Enterovirus in Taiwan

The current research aims to employ data mining technique to extract knowledge for Taiwans epidemic diseases open data including dengue, influenza, and enterovirus. Predication model using classification-oriented extraction mechanism to derive decision rules is considered. Predication accuracy and simplicity of decision tree are presented and compared. The data sources mainly include open data, climate data, and Google trend data. The Google trend data is included in the predication model and in comparison with not included. Research findings derived from analysis of 70,915 dengue cases, 52,908 influenza cases, and 34,063 enterovirus cases are obtained. implications and suggestions are also addressed.

Facilitating Active Multidimensional Association Mining with User Preference Ontology

Multidimensional association mining from data warehouse has become a knowledge discovery paradigm because it provides more specific conditional settings for target mining data, thus can generate rules more close to users’ needs. Yet data warehouse is subject to change by time or the modifications of business rules. Users might not know this change and reinitiate mining queries, which elicits the necessity of an active mining mechanism to bring new knowledge to users dynamically. In this paper, we propose an active multidimensional association mining system framework that incorporates with the user preference ontology that exploits frequent and representative queries. With the assistance of the user preference ontology and its association with user profile, the proposed system can facilitate active mining mechanism, allowing distribution of the renewal mining results to the specific users automatically.