Hasil Pencarian  ::  Simpan CSV :: Kembali

Abstrak :
Penelitian yang dilakukan pada laporan ini ditujukan untuk membuat suatu sistem yang dapat melakukan monitoring performa Wi-Fi, sehingga bila ada salah satu jaringan yang memiliki masalah dapat ditindak lanjuti dengan cepat untuk dianalisis lebih lanjut. Dalam penelitian ini juga ditambahkan peerbandingan kinerja machine learning untuk melakukan prediksi tentang bagaimana traffic wifi dapat berjalan ke depannya. Model machine learning yang dipakai pada penelitian ini adalah linear regression, Extreme Gradient Boost XGB regression, dan Light Gradient Boosting Machine (LGBM) regression. XGB dan LGBM merupakan pengembangan dari Gradient Boosting Decision Tree (GBDT). LGBM menerapkan dua Teknik yaitu Gradient-based One-Side Sampling (GOSS) dan Exclusive Features Bundling (EFB). GOSS berguna untuk mengecualikan gradien data kecil, sedangkan EFB berguna untuk memilih fitur eksklusif dengan tujuan pengurangan fitur. Peneliti menggunakan metrik evaluasi untuk mengetahui akurasi prediksi masing-masing model. Dari penelitian yang dilakukan metode machine learning LGBM lebih baik 3,09 % dari XGB regression dan 16,57 % lebih baik dari linear regression. ......This research in this report is aimed to create a system that can monitor WiFi performance, so that if one of the networks has problems it can be followed up quickly for further analysis. This research also add machine learning performance comparison to make predictions about how the WiFi traffic run in the future. The machine learning models used in this study are linear regression, Extreme Gradient Boost XGB regression, and Light Gradient Boosting Machine (LGBM) regression. LGBM applies two techniques namely Gradient-based One-Side Sampling (GOSS) and Exclusive Features Bundling (EFB). GOSS is useful for excluding small data gradients, whereas EFB is useful for selecting exclusive features with the goal of feature reduction. Researchers use evaluation metrics to determine the prediction accuracy of each model. From research conducted the LGBM machine learning method is 3.09% better than XGB regression and 16.57% better than linear regression.

Abstrak :
Contents : - The 1999 xDSL Comprehensive Report - About the IEC - University Program - Consortium-Affiliated Universities - Official Media Sponsors - Table of Contents - Table of Contents by Author - ACRONYM GUIDE - Foreword - ADSL: Prospects and Possibilities - Investing in Managing the Last Mile - xDSL Opportunities for PC Vendors - The Business of Delivering High-Speed Data To Multidwelling Units - Breaking the xDSL Profitability Barrier - An Overview of xDSL Products - The Last-Mile Infrastructure:Co-location Realities - ADSL Business: Key Applications That Drive the Business Case - CuNet Deployment: A CLEC Perspective - xDSL Early Deployment Results:Valley Telephone Cooperative - Using a Multistage Subtended DSLAM for Cost-Effective Deployment of ADSL Networks - Deploying xDSL on the Private Loop - Early Deployment Results and DSL Market Acceptance - DSL Data: Competitive local Exchange Carrier Analysis - Early Deployment Results of xDSL - VDSI Performance: Mixed-Services Deployment - Cable Modems: A Broadband Solution to the Home - Alternative Broadband Access Technologies - The Bright Future of ATM and DSL Access Networks - Etherloop - Network Architectures - Directions in Digital Subscriber line Silicon Connectivity Solutions - xDSL Deployment Strategy Based on a Migration from an IP-Based to a Cell-Based Architecture - Technical Competition and the Impact on xDSL - Customer Premises Network Architecture for DSL Remote - Turning Copper into Gold - xDSL Technology - xDSL: Semiconductor Trends and Opportunities - xDSL Backbone Solutions - ADSL Operations and Network Management - Architecture and Technology Choices - The Evolution of Copper Access Transmission from Voiceband Modems to DSl Technologies - xDSL Technology Tutorial - Supercharging Telecom Products with ADSL Technology - Are There Any lessons learned From ISDN? - A Solution Suite for the xDSL Provisioning Process - DSL System Integration - DSL Technology: Simply the Best Choice - Optimizing Bundled Services:The Local Access Provider's - The World Will Not Look the Same in the Year 2005 - Computer Industry Support of xDSl - Defining Digital Loops: Avoiding Remonopolization in a Digital World - Computer Industry Issues for Universal ADSL - Formal DSl Standards Progress - Delivering High-Speed Internet to the Mass Market - Government Policies Affecting xDSL Deployment - xDSL Standards