Publications-Theses
Article View/Open
Publication Export
-
Google ScholarTM
NCCU Library
Citation Infomation
Related Publications in TAIR
題名 結合分類與迴歸技術之電信網路弱訊區域偵測混合模型
A Hybrid Model for Weak Signal Area Detection in Telecom Networks Using Classification and Regression Techniques作者 張舜欣
Chang, Shun-Hsin貢獻者 謝佩璇
張舜欣
Chang, Shun-Hsin關鍵詞 機器學習
弱訊區偵測
分類技術
迴歸模型
網路優化
Machine Learning
Weak Signal Detection
Classification Techniques
Regression Models
Network Optimization日期 2024 上傳時間 4-Feb-2025 16:11:18 (UTC+8) 摘要 電信網路優化是提升網路效能和使用者體驗的關鍵過程。隨著行動數據流量的急劇增長，在高密度區域中，大量使用者同時連線，造成頻譜資源迅速被瓜分，使每個用戶可獲得的有效頻寬下降。再加上建物造成的訊號遮蔽，以及各類無線訊號彼此干擾，導致訊號品質與覆蓋範圍嚴重受限。因此，辨識和解決弱訊區成為網路優化的重要目標。弱訊區指的是在特定區域內，訊號強度低於可接受標準的區域，可能導致用戶通話中斷、數據傳輸速度緩慢或網路無法連接等問題。透過標記這些弱訊區，電信業者可以進行針對性的改善措施，如增加基站或小型基站的佈建，調整現有基站的發射功功率與天線傾角，或優化頻譜資源的分配。 本研究旨在探討如何建立一個混合型的機器學習模型框架，針對已知訊號地點，運用分群暨分類技術辨識弱訊區；對於缺乏明確訊號資料的地點，則透過迴歸模型預測該地點的訊號指標。分群暨分類模型的結果可為已知訊號的地點生成一張弱訊區的初步分佈地圖；迴歸模型的結果，進一步補全未知地點的訊號指標，並更新地圖。最終，將兩者的結果結合，生成一張完整的弱訊區分佈圖，用於網路優化和天線調整參考。 實驗結果顯示，採用隨機森林和梯度提升決策樹進行分類時，相較於支援向量機和人工神經網路，兩者在各情境下均表現特別突出。在預測訊號指標方面， K最近鄰迴歸、決策樹迴歸與隨機森林迴歸的R2分數均接近0.9甚至以上，而線性迴歸表現相對較差。進一步的散佈圖分析顯示，隨機森林迴歸的數據點分佈更為緊密，且與參考線的吻合度更高，表現出優異的數據模式捕捉能力與良好的泛化能力。這些實驗結果充分證明了本研究所提出的混合型機器學習模型框架在處理弱訊區辨識和訊號預測方面的有效性，對於電信網路之優化具有實質的技術支援。
With surging mobile data traffic in dense areas, identifying and addressing weak-signal regions is crucial for improving telecommunication networks. Such regions, where signal strength falls below standards, can cause dropped calls and slow data rates. This study proposes a hybrid machine learning framework that incorporates clustering and classification to identify weak-signal areas from known data points, and regression to predict signal values where data are lacking. Integrating these results yields a comprehensive weak-signal distribution map to guide network optimization and antenna adjustments. Experiments show that Random Forest and Gradient Boosting consistently outperform other classifiers, while K-Nearest Neighbors Regression, Decision Tree Regression, and Random Forest regressors achieve R² scores exceeding 0.9, outperforming Linear Regression. Random Forest Regression further demonstrates superior alignment of predicted values with actual measurements. These findings confirm the framework’s effectiveness in identifying and predicting weak-signal regions, providing valuable support for telecommunication network optimization.參考文獻 [1] 行動寬頻服務用戶數統計https://www.ncc.gov.tw/chinese/gradation.aspx?site_content_sn=3152 [2] 內政部人口相關統計（人口結構） https://www.moi.gov.tw/cl.aspx?n=3922 [3] Guo, Y., Yu, L., Wang, Q., Ji, T., Fang, Y., Wei-Kocsis, J., & Li, P. （2021）. Weak Signal Detection in 5G+ Systems: A Distributed Deep Learning Framework. Proceedings of the Twenty-Second International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing, 201–210. https://doi.org/10.1145/3466772.3467049 [4] Cheng, T., Liu, C., & Ding, W.（2019）. Weak Signal Detection Based on Deep Learning. Proceedings of the 2019 4th International Conference on Multimedia Systems and Signal Processing, 114–118. https://doi.org/10.1145/3330393.3330409 [5] Al-Thaedan, A., Shakir, Z., Mjhool, A. Y., Alsabah, R., Al-Sabbagh, A., Nembhard, F., & Salah, M. (2024). A machine learning framework for predicting downlink throughput in 4G-LTE/5G cellular networks. International Journal of Information Technology, 16(2), 651–657. https://doi.org/10.1007/s41870-023-01678-w [6] Hastie, T., Tibshirani, R., & Friedman, J.（2009）. The Elements of Statistical Learning: Data Mining, Inference, and Prediction （2nd ed.）. Springer. http://doi.org/10.1007/978-0-387-84858-7 [7] Altman, N. S.（1992）. An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression. The American Statistician, 46（3）, 175-185. http://doi.org/10.1080/00031305.1992.10475879 [8] Breiman, L., Friedman, J., Olshen, R.A., & Stone, C.J.（1984）. Classification and Regression Trees （1st ed.）. Chapman and Hall/CRC. https://doi.org/10.1201/9781315139470 [9] Breiman, L.（2001）. Random Forests. Machine Learning 45, 5–32. https://doi.org/10.1023/A:1010933404324 [10] James, G., Witten, D., Hastie, T., Tibshirani, R., & Taylor, J.（2023）. An introduction to statistical learning : with applications in Python. Springer International Publishing. https://doi.org/10.1007/978-3-031-38747-0 [11] Liu, K., Li, A., Lin, X., Mao, Z., & Zhang, W.（2024）. Empirical study on the performance of various machine learning models in predicting stock price movements as a binary classification task. Applied and Computational Engineering. 55. 129-144. https://doi.org/10.54254/2755-2721/55/20241403 [12] Hammood, L., Doğru, İ., & Kılıç, K.（2023）. Machine Learning-Based Adaptive Genetic Algorithm for Android Malware Detection in Auto-Driving Vehicles. Applied Sciences. 13. 5403. https://doi.org/10.3390/app13095403 [13] Hardman, M. F., Homkrajae, A., Eaton-Magaña, S., Breeding, C. M., Palke, A. C., & Sun, Z. (2024). Classification of Gem Materials Using Machine Learning. Gems & Gemology, 60(3), 306–329. https://doi.org/10.5741/GEMS.60.3.306 [14] Kartikasari, P., Utami, I. T., Suparti, S., & Rahman, S. D. F.（2024）. Breast Cancer Classification Using Support Vector Machine（SVM） and Light Gradient Boosting Machine（LightGBM） Models. Media Statistika, 16（2）, 182–193. https://doi.org/10.14710/medstat.16.2.182-193 [15] J M, S. L., & P, S.（2024）. Unveiling the potential of machine learning approaches in predicting the emergence of stroke at its onset: A predicting framework. Scientific Reports, 14, 20053. https://doi.org/10.1038/s41598-024-70354-1 [16] Rajayyan, S., & Mustafa, S. M. M.（2023）. Comparative Analysis of Performance Metrics for Machine Learning Classifiers with a Focus on Alzheimer’s Disease Data. Acta Informatica Pragensia, 12（1）, 54–70. https://doi.org/10.18267/j.aip.198 [17] Kumar, D., & Ahamad, F.（2024）. Opinion Extraction using Hybrid Learning Algorithm with Feature Set Optimization Approach. Journal of Electrical Systems. 20. 1266-1276. https://doi.org/10.52783/jes.3694 [18] Kumar, S., Choudhary, M.K. & Thomas, T.（2024） A hybrid technique to enhance the rainfall-runoff prediction of physical and data-driven model: a case study of Upper Narmada River Sub-basin, India. Scientific Reports 14, 26263. https://doi.org/10.1038/s41598-024-77655-5 [19] Liang, W., Luo, S., Zhao, G., & Wu, H.（2020）. Predicting Hard Rock Pillar Stability Using GBDT, XGBoost, and LightGBM Algorithms. Mathematics, 8（5）, 765. https://doi.org/10.3390/math8050765 [20] Patnaik, A., Anagnostou, D. E., Mishra, R., ChristodoulouCG., & Lyke, J. C.（2006）. Applications of neural networks in wireless communications. IEEE Antennas and Propagation Magazine, 46（3）, 130–137. https://doi.org/10.1109/MAP.2004.1374125 [21] RTR 欄位定義 https://www.netztest.at/en/OpenDataSpecification.html#response-2 [22] Pedregosa, F., Varoquaux, G., Gramfort, A., et al.（2011）. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12, 2825-2830. [23] RTR-NetzTest https://www.netztest.at/en/Opendata [24] Ookla（Speedtest by Ookla Global Fixed and Mobile Network Performance Maps） https://registry.opendata.aws/speedtest-global-performance/ [25] Powers, D., & Ailab.（2011）. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness & correlation. Journal of Machine Learning Technologies. 2（1）. https://doi.org/10.9735/2229-3981 [26] Campello R., Moulavi D., Zimek A., & Sander J.（2015）. Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection. ACM Trans. Knowl. Discov. Data 10, 1, Article 5, 51 pages. https://doi.org/10.1145/2733381 [27] Ester, M., Kriegel, H.-P., Sander, J., & Xu, X. (1996). A density-based algorithm for discovering clusters in large spatial databases with noise. Knowledge Discovery and Data Mining, 226–231. https://dl.acm.org/doi/10.5555/3001460.3001507 [28] Bishop, C. (2006). Pattern Recognition and Machine Learning. Journal of Electronic Imaging. 16(4):140-155. https://doi.org/10.1117/1.2819119 描述 碩士
國立政治大學
資訊科學系碩士在職專班
100971003資料來源 http://thesis.lib.nccu.edu.tw/record/#G0100971003 資料類型 thesis dc.contributor.advisor 謝佩璇 zh_TW dc.contributor.author (Authors) 張舜欣 zh_TW dc.contributor.author (Authors) Chang, Shun-Hsin en_US dc.creator (作者) 張舜欣 zh_TW dc.creator (作者) Chang, Shun-Hsin en_US dc.date (日期) 2024 en_US dc.date.accessioned 4-Feb-2025 16:11:18 (UTC+8) - dc.date.available 4-Feb-2025 16:11:18 (UTC+8) - dc.date.issued (上傳時間) 4-Feb-2025 16:11:18 (UTC+8) - dc.identifier (Other Identifiers) G0100971003 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/155514 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系碩士在職專班 zh_TW dc.description (描述) 100971003 zh_TW dc.description.abstract (摘要) 電信網路優化是提升網路效能和使用者體驗的關鍵過程。隨著行動數據流量的急劇增長，在高密度區域中，大量使用者同時連線，造成頻譜資源迅速被瓜分，使每個用戶可獲得的有效頻寬下降。再加上建物造成的訊號遮蔽，以及各類無線訊號彼此干擾，導致訊號品質與覆蓋範圍嚴重受限。因此，辨識和解決弱訊區成為網路優化的重要目標。弱訊區指的是在特定區域內，訊號強度低於可接受標準的區域，可能導致用戶通話中斷、數據傳輸速度緩慢或網路無法連接等問題。透過標記這些弱訊區，電信業者可以進行針對性的改善措施，如增加基站或小型基站的佈建，調整現有基站的發射功功率與天線傾角，或優化頻譜資源的分配。 本研究旨在探討如何建立一個混合型的機器學習模型框架，針對已知訊號地點，運用分群暨分類技術辨識弱訊區；對於缺乏明確訊號資料的地點，則透過迴歸模型預測該地點的訊號指標。分群暨分類模型的結果可為已知訊號的地點生成一張弱訊區的初步分佈地圖；迴歸模型的結果，進一步補全未知地點的訊號指標，並更新地圖。最終，將兩者的結果結合，生成一張完整的弱訊區分佈圖，用於網路優化和天線調整參考。 實驗結果顯示，採用隨機森林和梯度提升決策樹進行分類時，相較於支援向量機和人工神經網路，兩者在各情境下均表現特別突出。在預測訊號指標方面， K最近鄰迴歸、決策樹迴歸與隨機森林迴歸的R2分數均接近0.9甚至以上，而線性迴歸表現相對較差。進一步的散佈圖分析顯示，隨機森林迴歸的數據點分佈更為緊密，且與參考線的吻合度更高，表現出優異的數據模式捕捉能力與良好的泛化能力。這些實驗結果充分證明了本研究所提出的混合型機器學習模型框架在處理弱訊區辨識和訊號預測方面的有效性，對於電信網路之優化具有實質的技術支援。 zh_TW dc.description.abstract (摘要) With surging mobile data traffic in dense areas, identifying and addressing weak-signal regions is crucial for improving telecommunication networks. Such regions, where signal strength falls below standards, can cause dropped calls and slow data rates. This study proposes a hybrid machine learning framework that incorporates clustering and classification to identify weak-signal areas from known data points, and regression to predict signal values where data are lacking. Integrating these results yields a comprehensive weak-signal distribution map to guide network optimization and antenna adjustments. Experiments show that Random Forest and Gradient Boosting consistently outperform other classifiers, while K-Nearest Neighbors Regression, Decision Tree Regression, and Random Forest regressors achieve R² scores exceeding 0.9, outperforming Linear Regression. Random Forest Regression further demonstrates superior alignment of predicted values with actual measurements. These findings confirm the framework’s effectiveness in identifying and predicting weak-signal regions, providing valuable support for telecommunication network optimization. en_US dc.description.tableofcontents 第一章 緒論 7 1.1 研究背景與動機 7 1.2 研究目的 7 1.3 研究貢獻 8 1.4 研究流程 8 第二章 相關研究的技術內容 9 2.1 機器學習在電信網路下載速度預測中的應用 10 2.2 機器學習在各領域與醫學的應用與啟發 12 2.3 混合模型：本研究核心技術及其效益 14 2.4 機器學習的分群技術 16 2.5 機器學習的分類技術 18 2.5.1 支援向量機 18 2.5.2 隨機森林 19 2.5.3 梯度提升決策樹 19 2.5.4 人工神經網路 20 2.6 機器學習迴歸模型 21 2.6.1 線性迴歸 21 2.6.2 K最近鄰迴歸 21 2.6.3 決策樹迴歸 22 2.6.4 隨機森林迴歸 22 第三章 研究方法 23 3.1 系統架構設計 23 3.2 環境建置 29 3.3 資料來源 31 3.3.1 RTR-NetzTest 31 3.3.2 Ookla 31 3.4 資料處理 32 3.4.1 數據網格轉換經緯度座標 32 3.5 模型訓練方式 33 3.5.1 分群模型 33 3.5.2 分類模型 34 3.5.3 迴歸模型 37 3.6 模型評估及優化方式 41 3.6.1 分類評估方式 41 3.6.2 迴歸評估方式 42 3.6.3 優化方式 42 第四章 實驗過程與結果分析 44 4.1 實驗一：使用RTR-NETZTEST資料集（奧地利維也納） 46 4.1.1 分群標示弱訊標籤 46 4.1.2 分類辨識弱訊區 46 4.1.3 迴歸模型預測RSRP 52 4.2 實驗二：使用OOKLA資料集（美國波士頓） 54 4.2.1 分群標示弱訊標籤 54 4.2.2 分類辨識弱訊區 54 4.2.3 迴歸模型預測RSRP 58 4.3 實驗三：使用OOKLA資料集（臺灣臺北市） 60 4.3.1 分群標示弱訊標籤 60 4.3.2 分類辨識弱訊區 60 4.3.3 迴歸模型預測RSRP 64 4.4 實驗結果總結 68 第五章 結論與未來研究方向 69 第六章 參考文獻 71 zh_TW dc.format.extent 3955752 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0100971003 en_US dc.subject (關鍵詞) 機器學習 zh_TW dc.subject (關鍵詞) 弱訊區偵測 zh_TW dc.subject (關鍵詞) 分類技術 zh_TW dc.subject (關鍵詞) 迴歸模型 zh_TW dc.subject (關鍵詞) 網路優化 zh_TW dc.subject (關鍵詞) Machine Learning en_US dc.subject (關鍵詞) Weak Signal Detection en_US dc.subject (關鍵詞) Classification Techniques en_US dc.subject (關鍵詞) Regression Models en_US dc.subject (關鍵詞) Network Optimization en_US dc.title (題名) 結合分類與迴歸技術之電信網路弱訊區域偵測混合模型 zh_TW dc.title (題名) A Hybrid Model for Weak Signal Area Detection in Telecom Networks Using Classification and Regression Techniques en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] 行動寬頻服務用戶數統計https://www.ncc.gov.tw/chinese/gradation.aspx?site_content_sn=3152 [2] 內政部人口相關統計（人口結構） https://www.moi.gov.tw/cl.aspx?n=3922 [3] Guo, Y., Yu, L., Wang, Q., Ji, T., Fang, Y., Wei-Kocsis, J., & Li, P. （2021）. Weak Signal Detection in 5G+ Systems: A Distributed Deep Learning Framework. Proceedings of the Twenty-Second International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing, 201–210. https://doi.org/10.1145/3466772.3467049 [4] Cheng, T., Liu, C., & Ding, W.（2019）. Weak Signal Detection Based on Deep Learning. Proceedings of the 2019 4th International Conference on Multimedia Systems and Signal Processing, 114–118. https://doi.org/10.1145/3330393.3330409 [5] Al-Thaedan, A., Shakir, Z., Mjhool, A. Y., Alsabah, R., Al-Sabbagh, A., Nembhard, F., & Salah, M. (2024). A machine learning framework for predicting downlink throughput in 4G-LTE/5G cellular networks. International Journal of Information Technology, 16(2), 651–657. https://doi.org/10.1007/s41870-023-01678-w [6] Hastie, T., Tibshirani, R., & Friedman, J.（2009）. The Elements of Statistical Learning: Data Mining, Inference, and Prediction （2nd ed.）. Springer. http://doi.org/10.1007/978-0-387-84858-7 [7] Altman, N. S.（1992）. An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression. The American Statistician, 46（3）, 175-185. http://doi.org/10.1080/00031305.1992.10475879 [8] Breiman, L., Friedman, J., Olshen, R.A., & Stone, C.J.（1984）. Classification and Regression Trees （1st ed.）. Chapman and Hall/CRC. https://doi.org/10.1201/9781315139470 [9] Breiman, L.（2001）. Random Forests. Machine Learning 45, 5–32. https://doi.org/10.1023/A:1010933404324 [10] James, G., Witten, D., Hastie, T., Tibshirani, R., & Taylor, J.（2023）. An introduction to statistical learning : with applications in Python. Springer International Publishing. https://doi.org/10.1007/978-3-031-38747-0 [11] Liu, K., Li, A., Lin, X., Mao, Z., & Zhang, W.（2024）. Empirical study on the performance of various machine learning models in predicting stock price movements as a binary classification task. Applied and Computational Engineering. 55. 129-144. https://doi.org/10.54254/2755-2721/55/20241403 [12] Hammood, L., Doğru, İ., & Kılıç, K.（2023）. Machine Learning-Based Adaptive Genetic Algorithm for Android Malware Detection in Auto-Driving Vehicles. Applied Sciences. 13. 5403. https://doi.org/10.3390/app13095403 [13] Hardman, M. F., Homkrajae, A., Eaton-Magaña, S., Breeding, C. M., Palke, A. C., & Sun, Z. (2024). Classification of Gem Materials Using Machine Learning. Gems & Gemology, 60(3), 306–329. https://doi.org/10.5741/GEMS.60.3.306 [14] Kartikasari, P., Utami, I. T., Suparti, S., & Rahman, S. D. F.（2024）. Breast Cancer Classification Using Support Vector Machine（SVM） and Light Gradient Boosting Machine（LightGBM） Models. Media Statistika, 16（2）, 182–193. https://doi.org/10.14710/medstat.16.2.182-193 [15] J M, S. L., & P, S.（2024）. Unveiling the potential of machine learning approaches in predicting the emergence of stroke at its onset: A predicting framework. Scientific Reports, 14, 20053. https://doi.org/10.1038/s41598-024-70354-1 [16] Rajayyan, S., & Mustafa, S. M. M.（2023）. Comparative Analysis of Performance Metrics for Machine Learning Classifiers with a Focus on Alzheimer’s Disease Data. Acta Informatica Pragensia, 12（1）, 54–70. https://doi.org/10.18267/j.aip.198 [17] Kumar, D., & Ahamad, F.（2024）. Opinion Extraction using Hybrid Learning Algorithm with Feature Set Optimization Approach. Journal of Electrical Systems. 20. 1266-1276. https://doi.org/10.52783/jes.3694 [18] Kumar, S., Choudhary, M.K. & Thomas, T.（2024） A hybrid technique to enhance the rainfall-runoff prediction of physical and data-driven model: a case study of Upper Narmada River Sub-basin, India. Scientific Reports 14, 26263. https://doi.org/10.1038/s41598-024-77655-5 [19] Liang, W., Luo, S., Zhao, G., & Wu, H.（2020）. Predicting Hard Rock Pillar Stability Using GBDT, XGBoost, and LightGBM Algorithms. Mathematics, 8（5）, 765. https://doi.org/10.3390/math8050765 [20] Patnaik, A., Anagnostou, D. E., Mishra, R., ChristodoulouCG., & Lyke, J. C.（2006）. Applications of neural networks in wireless communications. IEEE Antennas and Propagation Magazine, 46（3）, 130–137. https://doi.org/10.1109/MAP.2004.1374125 [21] RTR 欄位定義 https://www.netztest.at/en/OpenDataSpecification.html#response-2 [22] Pedregosa, F., Varoquaux, G., Gramfort, A., et al.（2011）. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12, 2825-2830. [23] RTR-NetzTest https://www.netztest.at/en/Opendata [24] Ookla（Speedtest by Ookla Global Fixed and Mobile Network Performance Maps） https://registry.opendata.aws/speedtest-global-performance/ [25] Powers, D., & Ailab.（2011）. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness & correlation. Journal of Machine Learning Technologies. 2（1）. https://doi.org/10.9735/2229-3981 [26] Campello R., Moulavi D., Zimek A., & Sander J.（2015）. Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection. ACM Trans. Knowl. Discov. Data 10, 1, Article 5, 51 pages. https://doi.org/10.1145/2733381 [27] Ester, M., Kriegel, H.-P., Sander, J., & Xu, X. (1996). A density-based algorithm for discovering clusters in large spatial databases with noise. Knowledge Discovery and Data Mining, 226–231. https://dl.acm.org/doi/10.5555/3001460.3001507 [28] Bishop, C. (2006). Pattern Recognition and Machine Learning. Journal of Electronic Imaging. 16(4):140-155. https://doi.org/10.1117/1.2819119 zh_TW
