Chiun-Chieh Hsu

Using Topic Keyword Clusters for Automatic Document Clustering

Data clustering is a technique for grouping similar data items together for convenient understanding. Conventional data clustering methods, including agglomerative hierarchical clustering and partitional clustering algorithms, frequently perform unsatisfactorily for large text collections, since the computation complexities of the conventional data clustering methods increase very quickly with the number of data items. Poor clustering results degrade intelligent applications such as event tracking and information extraction. This paper presents an unsupervised document clustering method which identifies topic keyword clusters of the text corpus. The proposed method adopts a multi-stage process. First, an aggressive data cleaning approach is employed to reduce the noise in the free text and further identify the topic keywords in the documents. All extracted keywords are then grouped into topic keyword clusters using the k-nearest neighbor approach and the keyword clustering technique. Finally, all documents in the corpus are clustered based on the topic keyword clusters. The proposed method is assessed against conventional data clustering methods on a web news corpus. The experimental results show that the proposed method is an efficient and effective clustering approach.

Using topic keyword clusters for automatic document clustering

Data clustering is a technique for grouping similar data items together for convenient understanding. Conventional data clustering methods, including agglomerative hierarchical clustering and partitional clustering algorithms frequently perform unsatisfactorily for large text article collections, as well as the computation complexity of the conventional data clustering methods increase very quick with the number of data items. This paper presents a system for automatic document clustering by identifying topic keyword clusters of the text corpus. The proposed system adopts a multi-stage process. First, an aggressive data cleaning approach is employed to reduce the noise in the free text and further identify the topic keywords within the documents. All extracted keywords are then grouped into topic keyword clusters using the k-nearest neighbor graph approach and the keyword clustering function. Finally, all documents in the corpus are clustered based on the topic keyword clusters. The proposed method was assessed against conventional data clustering methods on a Web news collection, indicating that the proposed method is an efficient and effective clustering approach.

On the fault-tolerant embeddings of complete binary trees in the mesh interconnection networks

In this article, several schemes are proposed for embedding complete binary trees (CBT) into meshes. All of the proposed methods outperform those in the previous studies. First, a link congestion 1 embedding is achieved. Its expansion ratio is at the lowest level as we know now. Except for this superiority, it also provides another capability for fault tolerance to resist abnormal system faults, thus the embedding structure can be more guaranteed and the node utilization is raised further. Second, a link congestion 1 embedding with no-bending constraint is obtained. This scheme provides efficient CBT embedding for both the optical mesh and general mesh at the same time while keeping those good properties as the previous scheme. The last one is an optimal embedding which is applied to a 3D cubic mesh, where the node is almost fully utilized and its link congestion is 2.

An efficient algorithm for minimum feedback vertex sets in rotator graphs

For a rotator graph with n! nodes, Hsu and Lin [C.C. Hsu, H.R. Lin, H.C. Chang, K.K. Lin, Feedback Vertex Sets in Rotator Graphs, in: Lecture Notes in Comput. Sci., vol. 3984, 2006, pp. 158-164] first proposed an algorithm which constructed a feedback vertex set (FVS) with time complexity O(n^n^-^3). In addition, they found that the size of the FVS is n!/3, which was proved to be minimum. In this paper, we present an efficient algorithm which constructs an FVS for a rotator graph in O(n!) time and also obtains the minimum FVS size n!/3. In other words, this algorithm derives the optimal result with linear time complexity in terms of the number of nodes in the rotator graph.

On the fault-tolerant embedding of complete binary trees in the pancake graph interconnection network

Abstract This paper studies the problem of embedding complete binary trees ( CBT s) into an n -dimensional pancake graph ( P n ) with fault-tolerant capability. First, a new embedding scheme is developed for mapping a source CBT with height ∑ m=2 n ⌊ log m⌋ and dilation 2 onto the P n . This scheme not only embeds a CBT whose height is very close to the largest possible one, but also saves a lot of unused generators and generator products. Furthermore, these unused generators and generator products are used to recover faulty nodes and embed multiple CBT s. Maximally, near 2/3 nodes of the source CBT are allowed to be faulty at the same time and can be recovered by our scheme with dilation 4. Alternatively, a scheme which can embed a CBT with height ∑ m=2 n ⌊ log m⌋−1 is also given. In this case, if all nodes in the CBT are faulty, they can be recovered in the smallest number of recovery steps and only with dilation 4.

Feedback vertex sets in rotator graphs

This paper provides an algorithm for finding feedback vertex set in rotator graphs. Feedback vertex set is a subset of a graph whose removal causes an acyclic graph and is developed in various topologies of interconnected networks. In 1992, Corbett pioneered rotator graphs, whose interesting topological structures attract many researchers to publish relative papers in recent years. In this paper, we first develops feedback vertex set algorithm for rotator graphs. Our algorithm utilizes the technique of dynamic programming and generates a feedback vertex set of size n!/3 for a rotator graph of scale n, which contains n! nodes. The generated set size is proved to be minimum. Finding a minimum feedback vertex set is a NP-hard problem for general graphs. The time complexity of our algorithm, which finds a minimum feedback vertex set for a rotator graph of scale n, is proved to be O(n n-3 ).

Unsupervised document clustering based on keyword clusters

Due to the explosion growth of digital information, automatic document clustering or categorization has been an important research topic. Since document clustering has high dimension, the magnitude of the representation features will influence the efficiency and effect of the clustering and the precision of the clustering results. This paper presents an unsupervised document clustering method based on partitioning a weighted undirected graph. It initially discovers a set of tightly relevant keyword clusters that are disposed throughout the feature space of the collection of documents, and further clusters the documents into document clusters by using these keyword clusters. The experimental results show that the proposed approach can efficiently produce higher quality document clustering as compared with several well-known document clustering algorithms.

An Efficient Algorithm to Reduce the Inflations in Multi-Supertask Environment by Using a Transient Behavior Prediction Method

The supertask approach is a means of supporting non-migratory tasks in Pfair (proportionate-fair) scheduling systems. In this approach, tasks bound to the same processor are combined into a single server task, the supertask, which is scheduled as an ordinary Pfair task. When a supertask is scheduled, one of its component tasks is selected for execution. P. Holman et al. showed that component-task deadlines can be guaranteed by inflating each supertasks utilization. Their experimental results showed that the required inflation factors should be small in practice. Consequently, the average inflation produced by their rules is much greater than that actually required by the supertasks. First, we propose a notion of transient behavior prediction for supertasks, which predicts the latest possible finish time of subtasks that belong to supertasks. On this basis, we present an efficient schedulability algorithm for Pfair supertasks in which the deadlines of all component tasks can be guaranteed. We also propose a task merging process which combines the unschedulable supertasks with some Pfair tasks; hence, a new supertask can be scheduled in the system. Finally, we propose new reweighting functions that can be used when the previous two methods fail. Our functions produce a smaller inflation factor than previous work does. To demonstrate the efficacy of the supertasking approach, we present experimental evaluations of our algorithm, which decreases substantially the number of reweights and the size of inflation when there are many supertasks in the Pfair-scheduled systems.

Parallel directed graph algorithms on directional processor arrays with reconfigurable bus systems

Processors arrays with reconfigurable bus systems (abbreviated to PARBS) have been received a lot of attention in the last decade, and many undirected graph algorithms with constant time complexity have been proposed on PARBS. However, for a directed graph, it will be proved that connecting PARBS in the way proposed for undirected graphs generates paths which do not exist in the directed graph. This result may lead to incorrect solution for directed graph problems. Therefore, in this paper, a model named D-PARBS (Directional PARBS) is proposed for eliminating the non-existent paths. This model can be used to correctly identifying redundant arcs on directed graphs in constant time. Furthermore, by modifying the D-PARBS architecture, constant time algorithms with O(n3) processors are developed to solve topological sort, transitive closure, cyclic graph checking, and strongly connected component problems on directed graphs.

An efficient algorithm to reduce the inflations in multi-supertask environment by using a transient behavior prediction method

The supertask approach is a means of supporting non-migratory tasks in Pfair (proportionate-fair) scheduling systems. In this approach, tasks bound to the same processor are combined into a single server task, the supertask, which is scheduled as an ordinary Pfair task. When a supertask is scheduled, one of its component tasks is selected for execution. P. Holman et al. showed that component-task deadlines can be guaranteed by inflating each supertasks utilization. Their experimental results showed that the required inflation factors should be small in practice. Consequently, the average inflation produced by their rules is much greater than that actually required by the supertasks. First, we propose a notion of transient behavior prediction for supertasks, which predicts the latest possible finish time of subtasks that belong to supertasks. On this basis, we present an efficient schedulability algorithm for Pfair supertasks in which the deadlines of all component tasks can be guaranteed. We also propose a task merging process which combines the unschedulable supertasks with some Pfair tasks; hence, a new supertask can be scheduled in the system. Finally, we propose new reweighting functions that can be used when the previous two methods fail. Our functions produce a smaller inflation factor than previous work does. To demonstrate the efficacy of the supertasking approach, we present experimental evaluations of our algorithm, which decreases substantially the number of reweights and the size of inflation when there are many supertasks in the Pfair-scheduled systems.