English  |  正體中文  |  简体中文  |  Items with full text/Total items : 52359/87459 (60%)
Visitors : 9144433      Online Users : 288
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/106873

    Title: 軌跡模式探勘及應用之研究
    A study on trajectory pattern mining and applications
    Authors: 洪智傑
    Contributors: 國立交通大學
    Date: 2011-07
    Issue Date: 2016-06-02 09:09:51 (UTC+8)
    Abstract: 隨著行動裝置的普及,我們可以由多種的設備及來源收集使用者的軌跡資
    大規模收集軌跡資料的方式。我們提出了一個架構MDC 來降低資料傳輸量以及
    車輛的回報數目。在MDC 中,車輛利用模型來代表其感測到之讀數,並且利用
    在本篇論文的第三個主題中,我們提出了利用軌跡中的線索(Clue) 探勘軌跡
    結構。利用Editing distance 的概念,我們設計出比較使用者軌跡模式的距離函數。利用此距離函數,我們便可以推得出使用者的社群關係。
    With the pervasiveness of mobile devices, the location of users can be easily determined by either GPS devices or some positioning techniques. Moreover, wireless communication systems enable users to access various kinds of information from anywhere at any time. Nowadays, through smart phones or some portable devices, people could access location-based services or share their locations to their friends via social web sites, such as Google’s latitude service and
    Foursquare service. These phenomenons show that there will be an increasing amount of user trajectory data available. It is a challenge and interesting task to discover valuable knowledge. With knowledge mined, we could develop many novel applications from such a huge amount of user trajectory data.
    In this dissertation, we develop a series of research works for trajectory pattern mining and explore patterns mined for location-based social services. In our study, we present how to collect users’ trajectories first. Then, two kinds of mining algorithms are proposed. Finally, we develop a framework for mining location-
    based social community structures. We briefly introduce each work as follows:
    In the first work, we focused on the problem of data collection of trajectory data in a vehicular sensor network where every vehicles are equipped GPS and can communicate with each other in an ad-hoc manner. We proposed a frame-work MDC to reduce the amount of data transmission and the number of vehicles reporting their GPS data points. In MDC, model functions are derived to repre-
    sent the raw GPS data points such that only some coefficients that describe its
    movements are reported. An in-network aggregation mechanism determines a set
    of groups and for each group, only one vehicle needs to report traffic data, thereby
    further reducing the number of simultaneous connections.
    In the second work, we proposed a regression-based approach to mine user
    movement patterns from call detail records in a mobile computing system. Call
    detail records are viewed as random sample trajectory data, and user movement
    patterns are represented as movement functions. At first, the call detail records
    that capture frequent user movement behaviors are extracted. By exploring the
    spatiotemporal locality of movements, call detail records describing the similar
    behaviors are clustered. The movement functions can be represented by regression
    lines to best fit the location and time of call detail records.
    In the third work, we proposed an algorithm for discovering trajectory patterns
    by exploiting trajectory clues. In reality, there are many factors, such as sampling
    method, sampling frequency and device constraints, will affect the capability of
    original trajectory data capturing the actual movements. Even if trajectories can
    only reflect partial movements of a user, they reveal some trajectory clues about
    the moving behaviors hidden in trajectories. We first propose a clue-similarity to
    measure how much clue between two trajectories. Based on the clue-similarity,
    a graph-based clustering algorithm is proposed to group trajectories with similar
    moving behaviors into the same cluster. At last, for each group, the spatial and
    temporal information are aggregated into trajectory patterns.
    In the fourth work, we targeted at the problem of mining user communities
    in a location-based social network, where users in the same community have the
    similar movement behaviors. At the first, trajectory patterns of each user are orga-
    nized into a probabilistic suffix tree, which is viewed as a trajectory profile of each user. Inspired by the concept of the edit distance of two sequences, the distance
    function of two trees is proposed. Finally, in light of the distance of trees, a user
    communities in a location-based social network are found by clustering users with
    similar trajectory patterns.
    Appears in Collections:[資訊工程學系暨研究所] 學位論文

    Files in This Item:

    File Description SizeFormat
    摘要.pdf3937KbAdobe PDF77View/Open
    目錄.pdf1410KbAdobe PDF28View/Open

    All items in 機構典藏 are protected by copyright, with all rights reserved.

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - Feedback