scispace - formally typeset
Search or ask a question
Journal Article

Application of adaptive artificial immune algorithm to data mining

TL;DR: This adaptive artificial immune algorithm for clustering could achieve final network structure well matching the crude data feature and relieve the dependence on problem characteristic itself, because its thresholds were obtained from the dynamic immune network structure and were adapted well to the entire network structure during the process of evolution.
Abstract: In the artificial immune algorithm for clustering analysis,its suppression and stimulate thresholds determine cluster precision and network population scaleTheses thresholds adopt fixed value that is decided according to the problem characteristic itself and user's experienceHowever,this modus operandi results in narrow application situation and is dependent heavily on problem characteristic itselfTherefore,an adaptive artificial immune algorithm for clustering was proposedThis algorithm could achieve final network structure well matching the crude data feature and relieve the dependence on problem characteristic itself,because its thresholds were obtained from the dynamic immune network structure and were adapted well to the entire network structure during the process of evolutionExperimental results demonstrate its effectiveness
Citations
More filters
Proceedings ArticleDOI
23 Sep 2010
TL;DR: The simulated three-dimensional flight path by improved electric potential method and Artificial immune algorithm are compared and the result is: artificial immune algorithm is not only feasible for searching three dimensional path applications, but there is a great reduction of the computational time.
Abstract: Artificial immune algorithm (AIA) is used for offline as well as online path planning in this paper. Firstly, AIA is divided into three cases: 1. antibody is randomly selected to inoculate; 2. all antibodies are inoculated; 3. every antibody is inoculated, but inoculated location is selected at random. At the same time, we have simulated two kinds of threat uncertainties and one special occasion (no-fly-zone): changing the size of a certain kind of threat or increasing the number of unexpected threat. Based on above flight path planned, we adjust level path to meet maximum turning angle restriction, and vertical track to normal acceleration and curvature constraint by slope restriction algorithm and curvature smoothing algorithm making practical steering possible. And we also reduced the redundant path points (or frequent turning) to easy pilot manipulate and decrease the unsafe factors of flight. In addition, 3 orders B-spline curve is utilized for representing 3D path. At last, the simulated three-dimensional flight path by improved electric potential method and artificial immune algorithm are compared and the result is: artificial immune algorithm is not only feasible for searching three dimensional path applications, but there is a great reduction of the computational time.

9 citations


Cites background from "Application of adaptive artificial ..."

  • ...AIA is now widely used to plan path in the robotics community while not being attracted enough attention in airplane route planning....

    [...]

  • ...V. CONCLUSION In this paper, an improved AIA method is used to plan a collision-free path....

    [...]

  • ...At the same time, we compares the tracks by above methods with improved electric potential theory [18] (for brief, IPT) under the same conditions (see Figure 9, Table 4), and we have the result: the IPT’s computing time for 3.313 seconds which uses most time compared with the above-mentioned three cases of AIA....

    [...]

  • ...In recent years, although path planning using genetic algorithm (GA) has been studied in great detail, artificial immune algorithm (AIA) [9-14] (which is similar as GA to imitate natural biological behavior) has many advantages over genetic algorithm, for example, two genetic operators (crossover and mutation) of GA make it a certain degenerate....

    [...]

  • ...AIA has a capability of fast, random and global searching, and it can guarantee individual species diversity....

    [...]

Proceedings ArticleDOI
25 Dec 2009
TL;DR: In the algorithm, the clone and mutation mechanism of the artificial immune network is utilized to get the implicit ratings to reduce the data sparsity and improve the scalability of recommender system.
Abstract: With the increasingly expanding of E-commerce scale, some problems, such as data sparsity and scalability problems, caused by the traditional collaborative filtering technology which is widely used in the recommender systems of E-commerce are becoming more and more prominent At the same time, these problems decrease the recommender accuracy and influence the application effect of the recommender systems Aiming at these problems, this paper presents a collaborative filtering algorithm based on adaptive artificial immune network In the algorithm, the clone and mutation mechanism of the artificial immune network is utilized to get the implicit ratings to reduce the data sparsity The algorithm uses the clone suppression and network suppression to decrease the data dimension and improve the scalability of recommender system The experiment results indicate that the algorithm can improve the recommender accuracy

Cites methods from "Application of adaptive artificial ..."

  • ...This paper designs a CF algorithm based on adaptive aiNet([8,9,10]) utilizing the favorable scalability of aiNet to make up for the scalability shortage of the traditional CF algorithm based on users and decrease the data sparsity through the clone and mutation mechanism of the aiNet to generate the implicit rating ....

    [...]

  • ...for each iteration, do: a) Use (1) to compute the affinity between each antibody and the antigen, then the processing of normalization is applied to these affinity values according to the equation proposed in [8]....

    [...]