應用Python的sklearn機器學習預測臺北市老舊聚落防災策略

Using Python's sklearn machine learning to predict disaster prevention strategies for old settlements in Taipei City

陳智誠(Chih-cheng Chen) ； 杜敏誠(Min-cheng Tu) ； 施邦築(Ban-Jwu Shih)

《水保技術》 15卷4期 (2021/12) Pp. 54-69

摘要

本研究以臺北市32處老舊聚落為研究對象，針對32處過去巡勘監測持續列管、拆遷安置與解除列管三種決策結果作分析，利用包含決策樹等七種預測模式及分析變數因子選用5、8、9等3種方式納入預測分析，並將預測與實際決策作比較，將其正確率作分析，選擇最佳模型提供後續大數據分析，包含老舊聚落更細區域及列管139處坡地社區及所有坡地社區納入分析，作為後續坡地防災決策之參考。

關鍵字

決策樹 ； 隨機森林 ； 類神經網路

並列摘要

This study takes the 32 old settlements in Taipei City as the research object, and analyzes the three decision-making results of continuous pipeline, demolition, and de-manipulation of 32 past patrols and monitoring, using seven prediction models including decision trees and analysis variable factors Choose 5, 8, 9 and other 3 methods to be included in the forecast analysis, compare the forecast with the actual decision, analyze the correct rate, select the best model to provide follow-up big data analysis, including older settlements and more detailed areas and management 139 sloping communities and all sloping communities are included in the analysis as a reference for subsequent slope disaster prevention decisions.

並列關鍵字

DecisionTreeClassifier ； RandomForestClassifie ； Perceptron