Scholarlink

<p>Kidney cancer is among the leading causes of cancer deaths worldwide, thereby signifying the urgent need for newer computational techniques that improve drug discovery and rank therapeutic candidates more effectively. Modern drugs datasets often contain partial or complex structural information, thus underlining the importance of models capable to handle nonlinear interactions and structural variety. Motivated by such challenges, this study investigates the application of mathematical models based on graph theoretics along with physicochemical properties to develop Quantitative Structure–Property Relationship models for anti-kidney cancer drugs. The structural analysis were constructed for deeper understanding by using 2D molecular structures. Two different machine learning methods Ricker Wavelet Neural Network (RWNN) and Random Forest (RF) are used to build predictive models and their performance is carefully examined using conventional metrics: the Coefficient of Determination (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R^2\)</EquationSource> </InlineEquation>), Root Mean Square Error, Mean Absolute Error, and Mean Square Error. The study comparison revealed that the RWNN model achieved high predictive accuracy with <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(R^2\)</EquationSource> </InlineEquation> values up to 0.9964 for molar volume and 0.9946 for molar refractivity in RF model. These high-performing models outputs were then included into a Multi-Criteria Decision-Making framework utilising the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). By integrating predictive insights from both topological and physicochemical perspectives, the TOPSIS framework generated a robust, multifaceted ranking of candidate compounds, effectively prioritizing the most promising anti-kidney cancer drugs for further investigation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Computational intelligence-based QSPR modeling and decision-making for anti-kidney cancer drugs evaluation using machine learning algorithms

  • 期刊论文

  • 发表日期：

    2026年5月22日

查看全文
  • Wakeel Ahmed,
  • Maryam Alvi,
  • Shahid Zaman

摘要

Kidney cancer is among the leading causes of cancer deaths worldwide, thereby signifying the urgent need for newer computational techniques that improve drug discovery and rank therapeutic candidates more effectively. Modern drugs datasets often contain partial or complex structural information, thus underlining the importance of models capable to handle nonlinear interactions and structural variety. Motivated by such challenges, this study investigates the application of mathematical models based on graph theoretics along with physicochemical properties to develop Quantitative Structure–Property Relationship models for anti-kidney cancer drugs. The structural analysis were constructed for deeper understanding by using 2D molecular structures. Two different machine learning methods Ricker Wavelet Neural Network (RWNN) and Random Forest (RF) are used to build predictive models and their performance is carefully examined using conventional metrics: the Coefficient of Determination ( \(R^2\) ), Root Mean Square Error, Mean Absolute Error, and Mean Square Error. The study comparison revealed that the RWNN model achieved high predictive accuracy with \(R^2\) values up to 0.9964 for molar volume and 0.9946 for molar refractivity in RF model. These high-performing models outputs were then included into a Multi-Criteria Decision-Making framework utilising the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). By integrating predictive insights from both topological and physicochemical perspectives, the TOPSIS framework generated a robust, multifaceted ranking of candidate compounds, effectively prioritizing the most promising anti-kidney cancer drugs for further investigation.