Enhancing topological index of calcium chloride network through feature selection methods exploration

Area of Science:

• Inorganic Chemistry
• Computational Chemistry
• Materials Science

Background:

• Calcium chloride (CaCl2) is a versatile inorganic salt with widespread industrial and pharmaceutical applications.
• Its properties are crucial for various uses, including de-icing, dust control, and as a drying agent.
• Understanding its physio-chemical properties is essential for optimizing its applications.

Purpose of the Study:

• To compute topological indices, coindices, and reverse indices for calcium chloride (CaCl2).
• To apply machine learning strategies to identify the most suitable indices for predicting CaCl2's physio-chemical properties.
• To utilize regression techniques for predicting the heat of formation (HOF) of CaCl2 and identify influential features.

Main Methods:

• Calculation of various topological descriptors for CaCl2.
• Implementation of machine learning algorithms to correlate indices with properties.
• Application of regression models to predict heat of formation (HOF).

Main Results:

• Successful computation of topological indices, coindices, and reverse indices for CaCl2.
• Identification of a key set of indices predictive of CaCl2's physio-chemical properties through machine learning.
• Validation of predictive models for HOF, highlighting the most influential features.

Conclusions:

• Topological indices provide valuable insights into the physio-chemical properties of calcium chloride (CaCl2).
• Machine learning and regression techniques effectively predict properties like HOF, aiding in material characterization.
• This approach offers a pathway for understanding and optimizing the use of inorganic compounds.