We propose a heuristic method that generates a graph for order/degree problem. Target graphs of our heuristics have large order (> 4000) and diameter 3. We describe the observation of smaller graphs and basic structure of our heuristics. We also explain an evaluation function of each edge for efficient 2-opt local search. Using them, we found the best solutions for several graphs.

Econo-ESA is an economic scheme of the explicit semantic analysis (ESA). The scheme properly decreases the ESA index matrix dimensions to achieve faster process with similar results. This paper discusses index matrix dimensional reduction schemes of econo-ESA. We did experiments with several schemes: random selection, k-means clustering, norm-based clustering, densest, and sparsest schemes. Each resulted matrix had different element values and density. Our experimental results showed that the random selection scheme, which had the nearest density to the original index matrix, gave the best results. We thus conclude that the index matrix density is an additional feature which has to be considered in econo-ESA.

A critical problem with location fingerprinting is the considerable time and effort spent measuring received signal strengths at all location candidates to create the survey database. To reduce this cost, existing methods use a signal path loss model to interpolate part of the survey database from data actually measured at location candidates. However, the positioning accuracy can become degraded, especially in an indoor area delimited by walls. In this paper, we confirm the degradation in accuracy of the existing method through a preliminary experiment, and propose an accurate interpolation method for survey databases. In the proposed method, part of the survey data is interpolated using a path loss model containing wall attenuation. Furthermore, to confirm the effectiveness of the proposed method, we evaluate the location estimation performance with the interpolated survey database and also verify the interpolated data. The proposed method improves the positioning accuracy by 25% over that of the existing method.

We focus on the majority rule in crowdsourcing in this paper. One of the problems that concern crowdsourcing is how to reduce the number of votes for reducing the necessary cost and latency in obtaining results. In this paper, we attempt to reduce the number of votes while we obtain almost the same results with full votes. To accomplish this goal, we record the percentage of correct answers that each human worker provided. We report a preliminary study of making use of the records for reducing the number of votes with synthetic data and workers, and we discuss future directions.

Multipoint relaying is a key concept adopted in the optimized link state routing protocol (OLSR). It realizes efficient flooding of topology control (TC) messages. It simultaneously supports generating small topology information which is enough to find the shortest path for any pair of nodes. In this paper, we propose a method to find more efficient multipoint relay (MPR) set, in terms of reducing topology control traffic of OLSR. Our method shares MPRs between a node and its neighbor nodes, to reduce the MPR ratio. The MPR ratio is defined as the number of MPR nodes divided by the total number of nodes in the network. By decreasing the MPR ratio, the number of TC message generation is decreased. To overlap the MPR set with MPR sets of the neighbors, each node computes the degree of coverage for each of its MPRs, and appends the degrees into HELLO messages. When a node computes its MPR set, the node intends to share MPRs by referring neighbors' MPR sets and the degrees of their MPRs. Our MPR selection algorithm keeps the size of MPR set no more than that of the conventional MPR selection. Through simulations, the MPR ratio and the number of OLSR packets reduce by 6% and 4%, respectively, in moderately dense networks.

Semantic text similarity (STS) uses specific test collections as its performance evaluation measurement. The test collections consist of text pairs with the same meaning even though in different text form. The existence is scarce compared with information retrieval (IR) test collections. This paper investigates the possibility to reuse IR test collections for STS tasks. Text pairs are derived from the relevant pair of IR test collections. Latent semantic analysis (LSA) and explicit semantic analysis (ESA) evaluate Glasgow's test collections, which are provided by ACM SIGIR community. Jaccard index measures the lexical similarity. Recall metric measures retrievability of recycling test collection with two existing test collections, Microsoft research paraphrase corpus and Microsoft research video description corpus, as evaluation baselines. Evaluation yields a promising outcome; the evaluated test collections have low Jaccard index and their recall values between the two baselines.

It is desired to have a system that can recommend news articles according to interests of a user, which would change with time. In this paper, we attempt to improve a news recommendation system with supervised classification by integrating a clustering method into the system in order to adapt flexibly to the variety of interests of a user. To follow the changes of interests with time, we need to make not only the classifier but also the clustering module of the system be easily updatable. We construct clusters in the feature space from combining one-dimensional clusters to make the size of storage to hold for updating clusters be constant. We let the data distribution in each one-dimensional space be influenced by the clustering results from another one-dimensional space, thereby taking into account of data distribution in the original multiple dimensional space. Main contribution of this paper is to propose a method that can achieve both of the two goals: to improve the performance of a news recommendation system and to make the amount of data to hold for updating clusters be constant. Some experimental results are shown and the effective and weak points of our proposal are discussed.

Quick recovery from a failure is required essentially for distributed stream processing systems. We focus on single-node fail-stop failures occurred in high availability stream processing systems in this paper. One of high availability mechanisms is to provide a backup node for a processing node in the systems. We propose exploitation of backup nodes for reducing recovery cost in such an environment. We report some simulation results to show the effectiveness of our proposal.

In this paper, we introduce group participation support into decentralized peer-to-peer (P2P) Web caching systems. The support allows P2P Web caching systems to have more participators, and as a result, realizes highly scalable caching systems. An implementation design of systems is described in this paper. Some experimental results show that every node in our system can get better performance by the introduction of group participation support.

In this paper, we propose name resolution middleware that involves positional awareness to identify target hosts in wireless visible area communication (WVAC). WVAC is a wireless communication network that enables users to exchange information with nearby hosts, especially within the users' visible area, wherein the underlying network can be locally and temporarily constructed without the aid of any centralized administration. In such communication, the proposed middleware assists the user to intuitively identify a target host by presenting the relative positional relationship among the neighbor hosts. The middleware possesses the following characteristics: (1) it provides simple APIs to be easily utilized by various applications, (2) a dedicated server preliminarily located in the network is not required to operate it, and (3) it adopts a positioning method that could estimate the relative locations of neighbor hosts without any pre-configuration. Finally, we construct a prototype system and evaluate the middleware using this system. The results indicate that this middleware can sufficiently work even on resource-limited mobile devices.

In this paper, we propose three heuristics for high-efficiency selections of MPRs (multipoint relays) in link state QoS routing protocol: MIMS (maximizing integration metric selection), MQES (maximizing QoS-efficiency selection), and MCIS (minimizing cost-inefficiency selection). The basic approach of these selections is to determine a smaller set of MPRs that provide better QoS paths between any two nodes. The main objective in doing so is to maximize the QoS (quality of service) effect while limiting routing overhead for flooding control traffic. In addition, we evaluate the performance of the routing protocols with the proposed selections by simulation. The results indicate that the proposed methods achieve high-efficiency selection; as compared to an existing MPR selection taking into account the QoS requirements, MIMS reduces the maintenance cost by 30%, while the throughput of the resultant path is decreased by 13%, and MCIS reduces the cost by 21% without any decrease in the throughput.

In this paper we propose product line based component unification to reuse works required for component unification such as combined tests, WCET estimation, etc. and properties produced by component unification such as documents, binaries etc. We perform a case study of the personal destination screen with personal identification and find that component unification is worth to consider in product line engineering. Moreover, we present a sketch of an algorithm, which we expect to use in product line engineering tools, to exploit candidate components to be unified for strategic reuse along the product line with the feature model diagram of FORM product line engineering.