15:[["$","title","0",{"children":"```mermaid\ngraph TD\n subgraph Stage1 [階段 1：數據收集與預處理]\n ... | SciDraw Gallery"}],["$","meta","1",{"name":"description","content":"```mermaid\ngraph TD\n subgraph Stage1 [階段 1：數據收集與預處理]\n A[數據收集
材料數據表 1 & VOCs 數據表 2] --> B[輸入特徵]\n B --> B1[材料特徵
比表面積、孔徑、疏水性、極性]\n B --> B2[VOCs 特徵
分子量、logP、分子直徑、OAV 值]\n B --> C[輸出特徵
吸附匹配分數]\n C --> D[數據預處理]\n D --> D1[清理數值欄位
填補缺失值]\n D1 --> D2[篩選有效 VOCs
OAV 值 > 0]\n D2 --> D3[特徵工程
基於吸附理論計算匹配分數]\n end\n\n subgraph Stage2 [階段 2：模型訓練與評估]\n D3 --> E[數據集分割]\n E --> E1[80% 訓練集]\n E --> E2[20% 測試集]\n E1 --> F[特徵標準化
(僅適用於 SVM 和線性迴歸)]\n F --> G[機器學習模型]\n G --> G1[XGBoost]\n G --> G2[SVM]\n G --> G3[隨機森林]\n G --> G4[線性迴歸]\n G1 --> H[訓練與評估]\n G2 --> H\n G3 --> H\n G4 --> H\n H --> H1[5 折交叉驗證
(計算 R² 的平均值和標準差)]\n H1 --> H2[計算評估指標
訓練集 R²、測試集 R²、RMSE、MAE]\n H2 --> I[模型效能比較]\n I --> I1[依測試集 R² 排序]\n I1 --> J[最佳模型選擇
(基於測試集 R² 或交叉驗證效能)]\n end\n\n subgraph Stage3 [階段 3：特徵重要性分析與優化指導]\n J --> K[可解釋性分析]\n```"}],["$","link","2",{"rel":"manifest","href":"https://sci-draw.com/manifest.webmanifest","crossOrigin":"$undefined"}],["$","meta","3",{"name":"keywords","content":"ai for figure generation,AI scientific illustration generator,scientific data visualization,scientific illustration tool,create graphical abstract,graphical abstract maker,ai scientific figure generator,scientific figure maker,BioRender alternative,TOC graphics generator,publication-ready figures,SVG vector illustration"}],["$","link","4",{"rel":"canonical","href":"https://sci-draw.com/zh-TW/gallery/gal_1772606379088_uivz7sm"}],["$","link","5",{"rel":"alternate","hrefLang":"en","href":"https://sci-draw.com/gallery/gal_1772606379088_uivz7sm"}],["$","link","6",{"rel":"alternate","hrefLang":"zh-CN","href":"https://sci-draw.com/zh/gallery/gal_1772606379088_uivz7sm"}],["$","link","7",{"rel":"alternate","hrefLang":"ja","href":"https://sci-draw.com/ja/gallery/gal_1772606379088_uivz7sm"}],["$","link","8",{"rel":"alternate","hrefLang":"ko","href":"https://sci-draw.com/ko/gallery/gal_1772606379088_uivz7sm"}],["$","link","9",{"rel":"alternate","hrefLang":"de","href":"https://sci-draw.com/de/gallery/gal_1772606379088_uivz7sm"}],["$","link","10",{"rel":"alternate","hrefLang":"fr","href":"https://sci-draw.com/fr/gallery/gal_1772606379088_uivz7sm"}],["$","link","11",{"rel":"alternate","hrefLang":"es","href":"https://sci-draw.com/es/gallery/gal_1772606379088_uivz7sm"}],["$","link","12",{"rel":"alternate","hrefLang":"pt-BR","href":"https://sci-draw.com/pt-BR/gallery/gal_1772606379088_uivz7sm"}],["$","link","13",{"rel":"alternate","hrefLang":"zh-TW","href":"https://sci-draw.com/zh-TW/gallery/gal_1772606379088_uivz7sm"}],["$","link","14",{"rel":"alternate","hrefLang":"it","href":"https://sci-draw.com/it/gallery/gal_1772606379088_uivz7sm"}],["$","link","15",{"rel":"alternate","hrefLang":"ru","href":"https://sci-draw.com/ru/gallery/gal_1772606379088_uivz7sm"}],["$","meta","16",{"property":"og:title","content":"```mermaid\ngraph TD\n subgraph Stage1 [階段 1：數據收集與預處理]\n ... | SciDraw Gallery"}],["$","meta","17",{"property":"og:description","content":"```mermaid\ngraph TD\n subgraph Stage1 [階段 1：數據收集與預處理]\n A[數據收集
材料數據表 1 & VOCs 數據表 2] --> B[輸入特徵]\n B --> B1[材料特徵
比表面積、孔徑、疏水性、極性]\n B --> B2[VOCs 特徵
分子量、logP、分子直徑、OAV 值]\n B --> C[輸出特徵
吸附匹配分數]\n C --> D[數據預處理]\n D --> D1[清理數值欄位
填補缺失值]\n D1 --> D2[篩選有效 VOCs
OAV 值 > 0]\n D2 --> D3[特徵工程
基於吸附理論計算匹配分數]\n end\n\n subgraph Stage2 [階段 2：模型訓練與評估]\n D3 --> E[數據集分割]\n E --> E1[80% 訓練集]\n E --> E2[20% 測試集]\n E1 --> F[特徵標準化
(僅適用於 SVM 和線性迴歸)]\n F --> G[機器學習模型]\n G --> G1[XGBoost]\n G --> G2[SVM]\n G --> G3[隨機森林]\n G --> G4[線性迴歸]\n G1 --> H[訓練與評估]\n G2 --> H\n G3 --> H\n G4 --> H\n H --> H1[5 折交叉驗證
(計算 R² 的平均值和標準差)]\n H1 --> H2[計算評估指標
訓練集 R²、測試集 R²、RMSE、MAE]\n H2 --> I[模型效能比較]\n I --> I1[依測試集 R² 排序]\n I1 --> J[最佳模型選擇
(基於測試集 R² 或交叉驗證效能)]\n end\n\n subgraph Stage3 [階段 3：特徵重要性分析與優化指導]\n J --> K[可解釋性分析]\n```"}],"$L24","$L25","$L26","$L27","$L28","$L29","$L2a","$L2b","$L2c","$L2d","$L2e","$L2f","$L30","$L31"]