Clustering method for the construction of machine learning model with high predictive ability

Chemometrics and Intelligent Laboratory Systems, Journal Year: 2024, Volume and Issue: 246, P. 105084 - 105084

Published: Feb. 9, 2024

Language: Английский

---

Unsupervised Machine Learning-Based Image Recognition of Raw Infrared Spectra: Toward Chemist-like Chemical Structural Classification and Beyond Numerical Data

Journal of Chemical Information and Modeling, Journal Year: 2025, Volume and Issue: unknown

Published: March 19, 2025

Recent advances in artificial intelligence have significantly improved spectral data analysis. In this study, we used unsupervised machine learning to classify chemical compounds based on infrared (IR) images, without relying prior knowledge. The potential of for classification was demonstrated by extracting IR images from the Spectral Database Organic Compounds and converting them into 208,620-dimensional vector data. Hierarchical clustering 230 revealed distinct main clusters (A-G), each with specific subclusters exhibiting higher intracluster similarities. Despite challenges, including sensitivity deviations difficulty distinguishing delicate structures spectra low transparency fingerprint area, proposed image recognition approach exhibits good potential. Both principal component analysis k-means produced similar results. Furthermore, method high robustness noise. Tanimoto coefficient evaluate molecular similarity, providing valuable insights. However, some results deviated chemists' intuitions. study also highlighted that scaling composition formulas weights did not affect because high-dimensional features dominated process. A comparison obtained fingerprints, using adjusted Rand index as a metric, indicated provided better performance than numerical same resolution. Overall, demonstrates feasibility offers novel perspective complements traditional methods, although classifications may always align This has broader implications fields such drug discovery, materials science, automated analysis, where handling large, raw sets is essential.

Language: Английский

---

Discovering Net-Zero Chemical Processes and Pathways by Developing Circular Reaction Networks and Their Hierarchical Screening

Industrial & Engineering Chemistry Research, Journal Year: 2025, Volume and Issue: unknown

Published: May 20, 2025

Language: Английский

---

Optimal Feature Engineering in Machine Learning of Oxidative Coupling of Methane

Published: Jan. 1, 2024

Language: Английский

---