Failure prediction for hard disk drive using negative correlation and rank-level fusion

Name: Waraporn Tepin

LCSH: Hard Disks (Computer science)

LCSH: Fusion.

Othors: Data strorage technology -- Thesis.

Othors: King Mongkut's Institute of Technology Ladkrabang. -- Data strorage technology -- Thesis.

Abstract: This thesis proposes a Prediction of Customer Failure Modes in Hard Disk Drive (HDD) using Negative Correlation and Neural Networks Rank-Level Fusion applied on key parameters measured in the manufacturing process of a HDD. In our 4 classifiers, namely; Neural Networks (NN), Discriminant Analysis (DA), Bayesian Networks (BN), and Support Vector Machines (SVM) are applied to classify data obtained after simplification through Principal Component Analysis (PCA). The outcome of those classifiers is then evaluated via Negative Correlation (NC) to find the most negatively correlated for a two-class problem. The complexity and speed to choice of the classifier is reduced by NC; without NC, it is very likely that all combinations have to be evaluated before the best predictor can be constructed. Then all the 11 possible combinations from 4 classifiers are further aggregated using Rank-Level Fusion to evaluate the final prediction model, namely: Borda Count (BC), Logistic Regression (LR), and Neural Networks (NN). Finally, the combination of BN and SVM (NC approach) are aggregated using NN Rank-Level Fusion is selected to be the final model. The properly selected combination via NC and obtained through NN Fusion is an optimized prediction model. The resultant model is less complex but achieves comparable or superior accuracy for testing samples. This has saved computational process time and resources to derive the solution.

King Mongkut's Institute of Technology Ladkrabang. Central Library

Address: Bangkok

Email: Lifelong@kmitl.ac.th

Name: Yuttana Kitjaidure

Role: Thesis Advisor

Email : kkyuttan@kmitl.ac.th

Created: 2012

Modified: 2559-01-26

Issued: 2559-01-26

วิทยานิพนธ์/Thesis

application/pdf

CallNumber: Thesis W31f 2012

tha

DegreeName: Master of Engineering (Data Storage Technology)

Level: Master's Degree

Descipline: Data Storage Technology

Grantor: King Mongkut's Institute of Technology Ladkrabang

©copyrights สถาบันเทคโนโลยีพระจอมเกล้าเจ้าคุณทหารลาดกระบัง

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	Fulltext Waraporn Tepin.pdf	6.88 MB	29	2021-01-28 18:38:24

ใช้เวลา

0.019688 วินาที