Mon-Fong Jiang

A parallelized indexing method for large-scale case-based reasoning

Abstract Case-based reasoning (CBR) is a problem-solving methodology commonly seen in artificial intelligence. It can correctly take advantage of the situations and methods in former cases to find out suitable solutions for new problems. CBR must accurately retrieve similar prior cases for getting a good performance. In the past, many researchers proposed useful technologies to handle this problem. However, the performance of retrieving similar cases may be greatly influenced by the number of cases. In this paper, the performance issue of large-scale CBR is discussed and a parallelized indexing architecture is then proposed for efficiently retrieving similar cases in large-scale CBR. Several algorithms for implementing the proposed architecture are also described. Some experiments are made and the results show the efficiency of the proposed method.

A framework of features selection for the case-based reasoning

CBR is a problem solving technique that reuses past cases and experiences to find a solution to problems. A critical issue in case based reasoning is to select the correct and enough features to represent a case. For this reason, the analysis of cases and extraction of the necessary features to represent a case are highly recommended in building a CBR system. However, this task is difficult to carry out since such knowledge often cannot be successfully and exhaustively captured and represented. A framework of feature mining system for the case based reasoning including two phases is proposed. The techniques of feature selection, data analysis and machine learning can thus be effectively integrated. This will promote flexibility and expandability of the case based reasoning system.

Discovering Structure from Document Databases

Querying a database for document retrieval is often a process close to querying an answering expert system. In this work, we apply the knowledge discovery techniques to build an information retrieval system by regarding the structural document database as the expertise of the knowledge discovery. In order to elicit the knowledge embedded in the document structure, a new knowledge representation, named StructuralDocuments(SD), is defined and a transformation process which can transform the documents into a set of SDs is proposed. To evaluate the performance of our idea, we developed an intelligent information retrieval system which can help users to retrieve the required personnel regulations in Taiwan. In our experiments, it can be easily seen that the retrieval results using SD are better than traditional approaches.

Data types generalization for data mining algorithms

With the increasing use of database applications, mining interesting information from huge databases becomes of great concern and a variety of mining algorithms have been proposed in recent years. As we know, the data processed in data mining may be obtained from many sources in which different data types may be used. However, no algorithm can be applied to all applications due to the difficulty of fitting data types to the algorithm. The selection of an appropriate data mining algorithm is based not only on the goal of the application, but also the data fittability. Therefore, to transform the non-fitting data type into a target one is also important in data mining, but the work is often tedious or complex since a lot of data types exist in the real world. Merging the similar data types of a given selected mining algorithm into a generalized data type seems to be a good approach to reduce the transformation complexity. In this work, the data type fittability problem for six kinds of widely used data mining techniques is discussed and a data type generalization process, including merging and transforming phases is proposed. In the merging phase, the original data types of the data sources to be mined are first merged into the generalized ones. The transforming phase is then used to convert the generalized data types into the target ones for the selected mining algorithm. Using the data type generalization process, the user can select an appropriate mining algorithm just for the goal of the application without considering the data types.

Intelligent query agent for structural document databases

Abstract Querying a database for document retrieval is often a process close to querying an answering expert system. In this work, we apply the expert system techniques to the intelligent query agent establishment and regard the structural document database as the expertise which can be the objective of the knowledge acquisition. A new knowledge representation, named Structural Documents (SDs), is proposed to be the base of our model, and a transformation process from the raw data to the format of a database is applied. Based on the SDs, more suitable results could be inferenced rapidly by inference engine, and the flow of inference is also described. For implementation, an intelligent Chinese information retrieval system for personnel regulations by integrating knowledge-based and full-text searching techniques is proposed. In our experiments, the structural information of the documents can be acquired from the database using the knowledge extraction module. By observing the operating process of users, we found the query process of users are simplified.

A Similarity Indexing Method for the Data Warehousing - Bit-Wise Indexing Method

Data warehouse is an information provider that collects necessary data from individual source databases to support the analytical processing of decision-support functions. Recently, research about the indexing technologies of data warehousing has been proposed to help efficient on-line analytical processing (OLAP). In the past decades, some novel indexing technologies of data warehousing were proposed to retrieve the information precisely. However, the concept of similarity indexing technology in the increasingly larger data warehousing was seldom been discussed. In this paper, the performance issue of approximation indexing technology in the data warehousing is discussed and a new similarity indexing method, called bit-wise indexing method, and the corresponding efficient algorithms are proposed for retrieving the similar cases of a case-based reasoning system using a data warehouse to be the storage space. Some experiments are made for comparing the performance with two other methods and the results show the efficiency of the proposed method.

Run-time parallelization for partially parallel loops

In this paper, a run-time technique based on inspector-executor scheme is proposed to find available parallelism on loops in this paper. Our inspector can determine the wavefronts by building a DEF-USE table. Additionally, the process of inspector for finding the wavefronts, can be parallelized fully without any synchronization. Our executor can perform the loop iterations concurrently. For each wavefront in a loop, the auto-adapted function is used to get a tailored thread number rather than using fixed thread number for execution. Experimental results show that our new parallel inspector can handle complex data dependency patterns and reduce itself execution time obviously. Besides, the new partitioning strategy for executor can also improve the performance of run-time parallelization obviously.

Developing a sugar-cane breeding assistant system by a hybrid adaptive learning technique

The traditional sugar-cane breeding process depends on the determination of an experienced breeding researcher. Since the sugar-cane breeding problem in agriculture field is a complex one, the use of computer-aided methodology is very suitable to solve this problem. In this paper, we use the techniques of neural networks and genetic algorithms to construct a method in order to induce the sugar-cane cross model from the sugar-cane parent database established by the Taiwan Sugar Research Institute since 1990. The experimental results show that the correct percentage for testing is about 70%.

Genetic Algorithm for Extended Cell Assignment Problem in Wireless ATM Network

In this paper, we investigate the extended cell assignment problem which optimally assigns new and split cells in PCS (Personal Communication Service) to switches in a wireless ATM network. Given cells and switches in an ATM network (whose locations are fixed and known), the problem is assigning cells to switches in an optimum manner. We would like to do the assignment in as attempt to minimize a cost criterion. The cost has two components: one is the cost of handoffs that involve two switches, and the other is the cost of cabling. This problem is modeled as a complex integer programming problem and finding an optimal solution to this problem is NP-complete. A stochastic search method, based on a genetic approach is proposed to solve this problem. Simulation results showtha t genetic algorithm is robust for this problem.

A fuzzy inductive learning algorithm for parallel loop scheduling

The conventional symbolic learning algorithm can not infer data that contains fuzzy information. In the past few years, we have designed a parallel loop scheduling method called KPLS based upon a knowledge based approach, that chooses an appropriate schedule for a different loop to assign loop iterations to a multiprocessor system for achieving high speedup rates. Unfortunately, we found that the attributes that were applied in KPLS contain some fuzzy information, which are inapplicable to the traditional symbolic learning strategy for inferring some concept descriptions. In this paper, we apply a fuzzy set concept to an AQR learning algorithm that is called FAQR. FAQR which can induce fuzzy linguistic rules from fuzzy instances is then proposed to solve the above parallel loop scheduling problem. Some promising inference rules have been found and applied to infer the choice of parallel loop scheduling. We apply FAQR to the IRIS flower classification problem. Experimental results show that our method yields high accuracy in both domains.