Scholarlink

Acetylcholinesterase enzyme, which is responsible for the breakdown of acetylcholine, is a key therapeutic target to cure the Alzheimer Disease(AD). Whenever this enzyme is blocked, there is more acetylcholine accessible in the synaptic knob for utilisation, improving cognition and memory. The objective of the current effort is to develop machine learning methods for differentiating among AChE inhibitors and non-inhibitors. Inhibiting AChE activity helps treat AD by raising acetylcholine levels, which facilitate better communication between brain cells. Dataset is collected using ChemBL database and the same contains around 5905 bioactivity molecules. The data contains 3 classes [active, inactive and intermediate] and is pre-processed for handling missing data and handling duplicate values and further used in generation of Lipinski’s descriptors or rules for applying Exploratory Data Analysis(EDA) and to get target feature for model building. Models are built using Random Forest regression technique which gives 87.8% accuracy while other algorithms which are applied are SVM regression and SVM classification which have 61.4% and 74.8% accuracy as model performance. Random Forest algorithm performs best comparatively among all other implemented algorithms.

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

Prediction of Anti-Cholinesterase Inhibitor for Alzheimer Disease Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pushpa B. Patil,
  • Shweta Patil,
  • Abubakar Mulla,
  • Pooja Dolli

摘要

Acetylcholinesterase enzyme, which is responsible for the breakdown of acetylcholine, is a key therapeutic target to cure the Alzheimer Disease(AD). Whenever this enzyme is blocked, there is more acetylcholine accessible in the synaptic knob for utilisation, improving cognition and memory. The objective of the current effort is to develop machine learning methods for differentiating among AChE inhibitors and non-inhibitors. Inhibiting AChE activity helps treat AD by raising acetylcholine levels, which facilitate better communication between brain cells. Dataset is collected using ChemBL database and the same contains around 5905 bioactivity molecules. The data contains 3 classes [active, inactive and intermediate] and is pre-processed for handling missing data and handling duplicate values and further used in generation of Lipinski’s descriptors or rules for applying Exploratory Data Analysis(EDA) and to get target feature for model building. Models are built using Random Forest regression technique which gives 87.8% accuracy while other algorithms which are applied are SVM regression and SVM classification which have 61.4% and 74.8% accuracy as model performance. Random Forest algorithm performs best comparatively among all other implemented algorithms.