Dissertation > Excellent graduate degree dissertation topics show

Oriented data flow partial exception outlier dynamic mining algorithm research and application

Author: GaoKe
Tutor: ShaoFengJing
School: Qingdao University
Course: Computer Software and Theory
Keywords: Outlier n threshold Data stream Data Mining Network Intrusion Detection System
CLC: TP311.13
Type: Master's thesis
Year: 2010
Downloads: 68
Quote: 0
Read: Download Dissertation

Abstract


The main problem of most of the data mining algorithm is to discover the \the observed values ??of the different mechanisms. Tasks found by the isolated point can be described as follows: given a collection constituted by the N data points or objects and expected number of isolated points n, found that compared with the remaining data is the most inconsistent, abnormal or significantly different first n objects. Outlier discovery is an important research direction in data mining, has been applied in many fields, especially in network intrusion detection, intrusion user behavior is significantly different from the average network user, based on this, we study outlier discovery algorithm, and the outlier analysis algorithm is applied to the network intrusion detection. The local anomaly outlier LOF algorithm can be found in different density-based outlier, given to each object a local outlier factor to indicate the extent of the abnormal, closer to the isolated nature of the definition of the point. In recent years, a new data applications are received extensive attention, enduring relationship no longer save the data in these applications, but the large-scale, continuous, rapid, time-varying data stream. Continue to arrive in the reality-Fi record belongs to the streaming data, while the original static LOF algorithm is not suitable for real-time data stream mining due to the high complexity and can not adapt to changes in the data stream, the paper focuses on the data stream environment accurately identification and found isolated point, a number of isolated points n threshold automatically adjusts to the local anomaly outlier mining algorithm: n-IncLOF algorithm dynamically. The algorithm not only retains the advantages of the original static LOF algorithm can meet the flow quickly to changing data requirements of real-time continuous query. n-IncLOF based local anomaly LOF isolated point found algorithm, considering the unevenness of the isolated points in the stream data occurs, the adjustment function of the number of isolated-point threshold n, analysis of the algorithm during the insertion of the data points, delete and modify and other processes is given the n-IncLOF algorithm description and analysis of the complexity of the algorithm. Based the n-IncLOF algorithm developed abnormal network intrusion detection system. Designed and implemented using n-IncLOF algorithms do the detection engine characteristic properties of the mixture of host and network-based real-time network intrusion the abnormal detection system OutlierDIDS and KDD CUP99 data flow anomaly detection experiment proved the n-IncLOF algorithm through the system of isolated points The destination changes have strong adaptive capacity, compared to the original algorithm not only significantly improve the detection rate also reduces the rate of false positives, but also proved OutlierDIDS system to meet the adaptability and real-time network intrusion detection system on the effectiveness of the requirements.

Related Dissertations

  1. Resrarch on Multisignatures and Multisigncryptions from Identity-based,TN918.1
  2. Research on Spread Spectrum Code Acquisition of Double Threshold Based on Sliding Correlator,TN914.42
  3. Research of LOS Rate Estimation Method for Strapdown Imaging Seeker,TJ765
  4. A Study on Healthcare Product Marketing Based on Data Mining Technology,F426.72
  5. Bing- thick academic thought and clinical experience and empirical studies apply to turtle soups treatment of chronic kidney disease,R249.2
  6. Research on Fuzzy C-Mean Clustering Algorithm Based on Particle Swarm Optimization and Shuffled Frog Leaping Algorithm,TP18
  7. Based on data mining research tax audit case selection,F812.42
  8. Research on Application of Data Mining Technology in Degree of Satisfaction Analysis of Television Customers,TP311.13
  9. Research on Face Detection Based on Skin Color Segmentation and AdaBoost Algorithm,TP391.41
  10. An Algorithm on Clustering and Anomaly Detection for Multiple Data Streams,TP311.13
  11. Design and Research of a Tree Data Structure for Mass Data’s Comprehensive Evaluation,TP311.12
  12. The Threshold Voltage Model and Series Resistance Extraction Method of Polycrystalline Silicon Thin Film Transistors,TN321.5
  13. The Research and System Design of Assessment Model for Network Learning Based on Intelligent Computation,TP18
  14. Outlier Detection Techniques on Uncertain Moving Objects,TP311.13
  15. Research on Several Machine Learning Methods and Their Applications in Video-Based Fingerprint Verification,TP391.41
  16. College-oriented Library Management and Intelligent Analysis System,TP311.13
  17. Building Threshold GARCH Models of Time Series Based on Genetic Algorithms and Its Study and Application,F832.51
  18. Dynamical Multi-threshold Template Created and Application on DSP Image Processing,TP391.41
  19. Rules Extraction from Artificial Neural Networks for Classification Based Improvedant Colony Algorithm,TP183
  20. Index Evaluation and Research of Diabetes Nutrition Dietary System,R587.2
  21. Desiging and Realizing the WEB System of Shandong Shenghan College Library,TP311.52

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems
© 2012 www.DissertationTopic.Net  Mobile