Machine Learning Approaches for Predicting Song Popularity: A Case Study in Music Analytics DOI Open Access

Dr.A K Velmuruga

INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT, Journal Year: 2023, Volume and Issue: 07(11), P. 1 - 11

Published: Nov. 1, 2023

Comprehending the aspects that impact song popularity has become crucial in always changing music industry. This study explores field of predictive modeling using cutting-edge algorithms XGBoost and LightGBM. Predictive models developed by a large dataset includes variety musical variables, such as duration, tempo, lyrical content, release year. To improve models' capacity, approach extensive work. provide thorough assessment algorithms' performance, is divided into training testing sets. Additionally, effectiveness LightGBM forecasting evaluated comparison analysis. increase prediction accuracy, hyperparameter optimization methods—specifically, Optuna—are used to fine-tune them. In addition, looks at feature importance, illuminating elements that, eyes each algorithm, greatly add its appeal. Using rigorous cross-validation approach, are validated, their generalization capabilities shown. The performance metrics, which comprehensive picture predicted include mean absolute error, squared median R-squared. By providing comparative analysis two well-known machine learning methods for popularity, this paper advances rapidly developing analytics. results offer significant perspectives professionals data scientists who looking efficient approaches forecast across various genres. Keywords — Music Popularity Prediction; Machine Learning; XGBoost; LightGBM; Modeling; Feature Engineering; Hyperparameter Optimization; Data Analytics; Comparative Analysis; Song Characteristics; Genre Classification; Ensemble Models; Cross-Validation; Optuna; Preprocessing; Importance; Regression; Analytics.

Language: Английский

Computational screening for prediction of co-crystals: method comparison and experimental validation DOI
Fateme Molajafari, Tianrui Li, Mehrnaz Abbasichaleshtori

et al.

CrystEngComm, Journal Year: 2024, Volume and Issue: 26(11), P. 1620 - 1636

Published: Jan. 1, 2024

COSMO-RS and machine learning-based models can reduce the cost of screening identifying crystal coformers, facilitating discovery new cocrystals.

Language: Английский

Citations

13

Emerging Landscape of Computational Modeling in Pharmaceutical Development DOI
Yuriy A. Abramov, Guangxu Sun, Qun Zeng

et al.

Journal of Chemical Information and Modeling, Journal Year: 2022, Volume and Issue: 62(5), P. 1160 - 1171

Published: Feb. 28, 2022

Computational chemistry applications have become an integral part of the drug discovery workflow over past 35 years. However, computational modeling in support development has remained a relatively uncharted territory for significant both academic and industrial communities. This review considers workflows three key components preclinical clinical development, namely, process chemistry, analytical research as well product formulation development. An overview each step respective is presented. Additionally, context solid form design, special consideration given to modern physics-based virtual screening methods. covers rational approaches polymorph, coformer, counterion, solvent selection design.

Language: Английский

Citations

37

Efficient Screening of Coformers for Active Pharmaceutical Ingredient Cocrystallization DOI Creative Commons
Isaac J. Sugden, Doris E. Braun, David Bowskill

et al.

Crystal Growth & Design, Journal Year: 2022, Volume and Issue: 22(7), P. 4513 - 4527

Published: June 15, 2022

Controlling the physical properties of solid forms for active pharmaceutical ingredients (APIs) through cocrystallization is an important part drug product development. However, it difficult to know a priori which coformers will form cocrystals with given API, and current state-of-the-art cocrystal discovery involves expensive, time-consuming, and, at early stages development, API material-limited experimental screen. We propose systematic, high-throughput computational approach primarily aimed identifying API/coformer pairs that are unlikely lead experimentally observable can therefore be eliminated only brief check, from any investigation. On basis well-established crystal structure prediction (CSP) methodology, proposed derives its efficiency by not requiring expensive quantum mechanical calculations beyond those already performed CSP investigation neat itself. The assumptions tested on 30 potential 1:1 multicomponent systems (cocrystals solvate) involving 3 9 one solvent. This complemented detailed all pairs, led five new (three API-coformer combinations, polymorphic example, different stoichiometries) cis-aconitic acid polymorph. indicates that, some APIs, significant proportion could investigated thereby saving considerable effort.

Language: Английский

Citations

33

Virtual Screening, Structural Analysis, and Formation Thermodynamics of Carbamazepine Cocrystals DOI Creative Commons
Artem O. Surov, Anna G. Ramazanova, Alexander P. Voronin

et al.

Pharmaceutics, Journal Year: 2023, Volume and Issue: 15(3), P. 836 - 836

Published: March 3, 2023

In this study, the existing set of carbamazepine (CBZ) cocrystals was extended through successful combination drug with positional isomers acetamidobenzoic acid. The structural and energetic features CBZ 3- 4-acetamidobenzoic acids were elucidated via single-crystal X-ray diffraction followed by QTAIMC analysis. ability three fundamentally different virtual screening methods to predict correct cocrystallization outcome for assessed based on new experimental results obtained in study data available literature. It found that hydrogen bond propensity model performed worst distinguishing positive negative experiments 87 coformers, attaining an accuracy value lower than random guessing. method utilizes molecular electrostatic potential maps machine learning approach named CCGNet exhibited comparable terms prediction metrics, albeit latter resulted superior specificity overall while requiring no time-consuming DFT computations. addition, formation thermodynamic parameters newly evaluated using temperature dependences Gibbs energy. reactions between selected coformers be enthalpy-driven, entropy being statistically from zero. observed difference dissolution behavior aqueous media thought caused variations their stability.

Language: Английский

Citations

22

Advanced Feature Analysis for Enhancing Cocrystal Prediction DOI Creative Commons

Alessandro Cossard,

Chiara Sabena,

Gianluca Bianchini

et al.

Chemometrics and Intelligent Laboratory Systems, Journal Year: 2025, Volume and Issue: unknown, P. 105318 - 105318

Published: Jan. 1, 2025

Language: Английский

Citations

1

Improving the Physicochemical and Biopharmaceutical Properties of Active Pharmaceutical Ingredients Derived from Traditional Chinese Medicine through Cocrystal Engineering DOI Creative Commons
Danyingzi Guan, Bianfei Xuan,

Chengguang Wang

et al.

Pharmaceutics, Journal Year: 2021, Volume and Issue: 13(12), P. 2160 - 2160

Published: Dec. 15, 2021

Active pharmaceutical ingredients (APIs) extracted and isolated from traditional Chinese medicines (TCMs) are of interest for drug development due to their wide range biological activities. However, the overwhelming majority APIs in TCMs (T-APIs), including flavonoids, terpenoids, alkaloids phenolic acids, limited by poor physicochemical biopharmaceutical properties, such as solubility, dissolution performance, stability tabletability development. Cocrystallization these T-APIs with coformers offers unique advantages modulate properties drugs without compromising therapeutic benefits non-covalent interactions. This review provides a comprehensive overview current challenges, applications, future directions T-API cocrystals, cocrystal designs, preparation methods, modifications corresponding mechanisms properties. Moreover, variety studies presented elucidate relationship between crystal structures cocrystals resulting along underlying mechanism changes. It is believed that understanding engineering could contribute more bioactive natural compounds into new drugs.

Language: Английский

Citations

35

How Many Cocrystals Are We Missing? Assessing Two Crystal Engineering Approaches to Pharmaceutical Cocrystal Screening DOI
Chiara Cappuccino,

David F. Cusack,

James L. Flanagan

et al.

Crystal Growth & Design, Journal Year: 2022, Volume and Issue: 22(2), P. 1390 - 1397

Published: Jan. 10, 2022

Drug development may include extensive screening for crystalline forms of active pharmaceutical ingredients. Crystal engineering aims to apply supramolecular knowledge simplify such a task. The failure strategy result in overlooking potentially interesting compounds. Here, the advantages knowledge-based approach is compared systematic crystallization cocrystals. This work indicates that simply based on known synthons and their relative frequency as reported database effective random exercise, missing 25% successful cocrystallization. Readily available computational methods perform better, enabling identification all observed cocrystals with reduction 24% experimental attempts.

Language: Английский

Citations

29

Cocrystal Synthesis through Crystal Structure Prediction DOI
Yuriy A. Abramov, Luca Iuzzolino,

Yingdi Jin

et al.

Molecular Pharmaceutics, Journal Year: 2023, Volume and Issue: 20(7), P. 3380 - 3392

Published: June 6, 2023

Crystal structure prediction (CSP) is an invaluable tool in the pharmaceutical industry because it allows to predict all possible crystalline solid forms of small-molecule active ingredients. We have used a CSP-based cocrystal method rank ten potential coformers by energy cocrystallization reaction with antiviral drug candidate, MK-8876, and triol process intermediate, 2-ethynylglyclerol. For was performed retrospectively successfully predicted maleic acid as most likely be observed. The known form two different cocrystals 1,4-diazabicyclo[2.2.2]octane (DABCO), but larger landscape desired. screening triol-DABCO one, while triol-l-proline two. Computational finite-temperature corrections enabled determination relative crystallization propensities stoichiometries polymorphs free-energy landscape. obtained during subsequent targeted experiments found exhibit improved melting point deliquescence behavior over triol-free acid, which could considered alternative synthesis islatravir.

Language: Английский

Citations

15

In silico co-crystal design: Assessment of the latest advances DOI
Carolina von Eßen,

David Luedeker

Drug Discovery Today, Journal Year: 2023, Volume and Issue: 28(11), P. 103763 - 103763

Published: Sept. 7, 2023

Language: Английский

Citations

11

Exploring the Cocrystal Landscape of Posaconazole by Combining High-Throughput Screening Experimentation with Computational Chemistry DOI Creative Commons
Matteo Guidetti,

Rolf Hilfiker,

Martin Kuentz

et al.

Crystal Growth & Design, Journal Year: 2022, Volume and Issue: 23(2), P. 842 - 852

Published: Dec. 23, 2022

The development of multicomponent crystal forms, such as cocrystals, represents a means to enhance the dissolution and absorption properties poorly water-soluble drug compounds. However, successful discovery new pharmaceutical cocrystals remains time- resource-consuming process. This study proposes use combined computational-experimental high-throughput approach tool accelerate improve efficiency cocrystal screening exemplified by posaconazole. First, we employed COSMOquick software preselect rank candidates (coformers). Second, crystallization experiments (HTCS) were conducted on selected coformers. HTCS results successfully reproduced liquid-assisted grinding reaction crystallization, ultimately leading synthesis thirteen posaconazole (7 anhydrous, 5 hydrates, 1 solvate). characterized PXRD, 1H NMR, Fourier transform-Raman, thermogravimetry-Fourier transform infrared spectroscopy, differential scanning calorimetry. In addition, prediction performance was compared that two alternative knowledge-based methods: molecular complementarity (MC) hydrogen bond propensity (HBP). Although HBP does not perform better than random guessing for this case study, both MC show good discriminatory ability, suggesting their potential virtual screening.

Language: Английский

Citations

16