(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Extended nearest centroid neighbor method with training set reduction for classification / Nordiana Mukahar

Extended nearest centroid neighbor method with training set reduction for classification_Nordiana Mukahar_E3_2020_MYMY
Pengeluaran kejat mensasarkan Kaizen dari segi kualiti, kos dan tempoh kitaran. Penyelesaian masalah yang mantap selalunya melibatkan pihak luar seperti vendor, khususnya dalam aspek penggunaan sumber dan pengetahuan teknologi. Semakan terhadap penerbitan “peer-review” semasa menunjukkan bahawa kajian akademik terhadap aktiviti sedemikian adalah terhad, terutamanya yang mempunyai unsur-unsur perkongsian risiko dan ganjaran yang jelas. Dalam premis ini, rangka kerja Perkongsian Risiko dan Ganjaran - Kaizen (VRRS-Kaizen) telah dicadangkan sebagai sistem preskriptif yang generik dan holistik bagi membimbing interaksi dengan vendor, dengan matlamat untuk memastikan amalan yang sistematik dan berkesan. Rancang-Laksana-Periksa-Tindakan menjadi asas kepada rangka kerja tersebut dan membentuk tahap-tahap pelaksanaan Kaizen. VRRS-Kaizen bermula dengan mengenalpasti keperluan untuk mendapatkan khidmat vendor bagi tujuan pembangunan tindakbalas oleh Pasukan Kaizen. Kaedah Kejat, pembuktian konsep, dan kaedah pemarkahan pelbagai kriteria telah digunakan untuk penilaian dalam rangka kerja berkenaan. Pengesahan rangka kerja pula dilakukan melalui tiga kajian kes di sebuah syarikat sistem pengukuran elektronik di Pulau Pinang. Skop kerja melibatkan penghapusan 100% kegagalan suhu dalaman yang tinggi bagi “peranti-di-bawah-ujian” (Kajian Kes Satu), pengurangan penggunaan elektrik sebanyak 60.9% serta pengurangan caj penyelenggaraan sebanyak 55.6% (Kajian Kes Dua), dan mitigasi caj pengangkutan yang tinggi untuk Pemasangan Pembungkusan 64A (Kajian Kes Tiga). Dengan perbezaan yang jelas antara setiap kes, setiap kes Berjaya menggunakan rangka kerja yang dibincangkan. Secara keseluruhan, pelaksanaan tersebut telah diterjemahkan ke dalam bentuk pulangan sebanyak RM 204,105.86 (antara 2017 hingga 2018), di mana 45.52% daripada jumlah itu telah dikongsikan dengan vendor-vendor yang terlibat sebagai perkongsian ganjaran kewangan. Dengan itu, objektif kajian telah dicapai. _______________________________________________________________________ Lean manufacturing seeks Kaizen in terms of quality, cost and cycle time. A robust problem-solving often extends to external parties such as vendors, to draw in their unique technology resources and knowledge. The perusal of contemporary peer-reviewed literature reveals limited academic investigation onto such form of partnership; particularly vendor engagement having elements of properly defined risk and reward sharing. In this premise, Vendor Risk and Reward Sharing – Kaizen (VRRS-Kaizen) framework was proposed as a generic and holistic prescriptive system to guide personnel to duly deal with vendors. The objective of the framework is to ensure systematic and effective practice. Plan-Do-Check-Act underpins the framework and dichotomises the relevant stages of Kaizen. VRRS-Kaizen commences with the identification by Kaizen Team for the need of calling in vendors for countermeasure development. Lean tools, proof of concept and multi-criteria scoring methods were used for assessments in the framework. Framework verification was performed through three case studies at an electronic measurement system company in Penang. Their scopes involve 100% elimination of device under test high internal temperature failures (Case Study One), reduction of high workstation electricity by 60.9% and maintenance charges by 55.6% (Case Study Two) and mitigation of high freight charges of Packaging Assembly 64A by 24% (Case Study Three). Evidently different in nature, these three cases have been successfully deployed following the framework. In total, these were translated into RM 204,105.86 in return (between 2017 to 2018), of which 45.52% was shared with vendors as financial reward sharing. The research objectives have been achieved.
Contributor(s):
Nordiana Mukahar - Author
Primary Item Type:
Thesis
Identifiers:
Accession Number : 875008821
Language:
English
Subject Keywords:
literature; Kaizen; holistic
Sponsor - Description:
Pusat Pengajian Kejuruteraan Elektrik & Elektronik -
First presented to the public:
6/1/2020
Original Publication Date:
7/23/2020
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 312
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2020-07-23 10:33:40.716
Submitter:
Mohamed Yunus Yusof

All Versions

Thumbnail Name Version Created Date
Extended nearest centroid neighbor method with training set reduction for classification / Nordiana Mukahar1 2020-07-23 10:33:40.716