แจ้งเอกสารไม่ครบถ้วน, ไม่ตรงกับชื่อเรื่อง หรือมีข้อผิดพลาดเกี่ยวกับเอกสาร ติดต่อที่นี่ ==>
หากไม่มีอีเมลผู้รับให้กรอก thailis-noc@uni.net.th
หากไม่มีอีเมลผู้รับให้กรอก thailis-noc@uni.net.th
Kittakorn Sriwanna. Graph clustering-based discretization technique : algorithms and applications. Doctoral Degree(Computer Engineering). Mae Fah Luang University . : Mae Fah Luang University , 2016.
Graph clustering-based discretization technique : algorithms and applications
Name: Kittakorn Sriwanna
Organization :
Mae Fah Luang University
keyword:
Graph clustering-based
LCSH:
Graph algorithms
Classification :.LCCS: QA166.245
LCSH:
Data mining
Abstract:
Discretization plays a major role in the data preprocessing technique used in
machine learning and data mining. Recent studies have been motivated to multivariate
discretization methods that consider the relation among attributes. The general goal
of the method is to obtain the discrete data that preserve as much information in the
original continuous data. However, most of them generate the final discrete data that
incomplete information, do not preserve the natural group of data, and usually require
more computation time. In addition to the class information, it may also be useful to
exploit natural data groups or clusters that normally exhibit a set of distinct properties
found in the data under examination. The current research strives to achieve this, which
has been lost in existing methods.
This graph clustering-based discretization approach present in this dissertation is
a novel technique that changes the data points as a graph representation. The proposed
approach is the first to address explicitly employ the natural group of data via
graph-based and discretize using graph clustering, on the basis of normalized
association in order to preserve as much as information. In particular, three new
methods of: Spiting (GraphS), Merging (GraphM), and Evolutionary (GraphE) based on
graph clustering-based discretization algorithm are proposed. The effectivenesses of the
proposed methods are empirically demonstrated over 30 standard datasets, 44 imbalance
datasets, 44 SMOTE datasets, and 10 microarray datasets. The proposed methods
compared to many state-of-the-art discretization algorithms using four classifiers. The
results reveal that the proposed methods can preserve the natural group of data and
prevents over generate a small interval. Consequently, the proposed methods obtain
higher classification performances over the existing discretization algorithms.
Mae Fah Luang University
Address:
CHIANG RAI
Email:
library@mfu.ac.th
Name: Natthakan Iam-On
Role:
Advisor
Created:
2016
Issued:
2560
Modified:
2020-04-29
Issued:
2560-09-06
วิทยานิพนธ์/Thesis
application/pdf
CallNumber:
Thesis QA166.245 K625g 2016
tha
DegreeName: Dotor of Philosophy
Level: Doctoral Degree
Descipline: Computer Engineering
Grantor: Mae Fah Luang University
©copyrights มหาวิทยาลัยแม่ฟ้าหลวง
RightsAccess:
ใช้เวลา
0.029285 วินาที
Kittakorn Sriwanna
| Title | Contributor | Type |
|---|---|---|
| Graph clustering-based discretization technique : algorithms and applications
มหาวิทยาลัยแม่ฟ้าหลวง Kittakorn Sriwanna | Asst. Prof. Natthakan Iam-On, Ph. D. | วิทยานิพนธ์/Thesis |
Natthakan Iam-On
| Title | Creator | Type and Date Create |
|---|---|---|
| Predicting progession duration of chronic kidney disease in patients with hypertension and diabetes
มหาวิทยาลัยแม่ฟ้าหลวง Asst. Prof. Natthakan Iam-On, Ph. D. | Warangkana Khannara | วิทยานิพนธ์/Thesis |
| Graph clustering-based discretization technique : algorithms and applications
มหาวิทยาลัยแม่ฟ้าหลวง Asst. Prof. Natthakan Iam-On, Ph. D. | Kittakorn Sriwanna | วิทยานิพนธ์/Thesis |
| Lmproved classification of adversarial network intrusion using ensemble clustering based data transformation
มหาวิทยาลัยแม่ฟ้าหลวง Natthakan Iam-on | Pongsakorn Tatongjai | วิทยานิพนธ์/Thesis |
| Image-based transient detection for GOTO sky survey
มหาวิทยาลัยแม่ฟ้าหลวง Natthakan Iam-on | Terry Cortez | วิทยานิพนธ์/Thesis |