The application of artificial intelligence system and regression methods based on the spectrophotometric method for fast simultaneous determination of naphazoline and antazoline in ophthalmic formulation

In this study, reliable and quick methods based on ultraviolet-visible spectrophotometry (UV–Vis) for the simultaneous determination of Naphazoline (NAP) and Antazoline (ANT) were proposed. Feed-forward back-propagation neural network (FFBP-NN) as an artificial intelligence system as well as partial...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Optik (Stuttgart) Ročník 203; s. 164010
Hlavní autori: Shahrokhi, Zeinab, Sohrabi, Mahmoud Reza, Nik, Saeed Mortazavi
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier GmbH 01.02.2020
Predmet:
Antazoline
Artificial intelligence
Naphazoline
Regression methods
Spectrophotometry
Naphazoline
Antazoline
Spectrophotometry
Artificial intelligence
Regression methods
ISSN:0030-4026, 1618-1336
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:In this study, reliable and quick methods based on ultraviolet-visible spectrophotometry (UV–Vis) for the simultaneous determination of Naphazoline (NAP) and Antazoline (ANT) were proposed. Feed-forward back-propagation neural network (FFBP-NN) as an artificial intelligence system as well as partial least squares (PLS) and principal component regression (PCR) as regression methods were used for this work. FFBP neural network with Levenberg–Marquardt (LM), scaled conjugate gradient (SCG), and gradient descent with momentum and adaptive learning rate backpropagation (GDX) algorithms with various layers and neurons were applied to select the best model. The lowest mean square error (MSE) related to the LM, SCG, and GDX algorithms was 3.44×10−14, 6.93×10-30- 1.13×10-7, 2.76×10-6, and 3.29×10-5, 9.63×10-3 for NAP and ANT, respectively. On the other hand, the mean recovery (%) and root mean square error (RMSE) of the PLS method were 100.03, 0.00032 and 100.03, 0.00326 for NAP and ANT, respectively. Also, the mean recovery and RMSE values corresponding to the PCR model were obtained 100.65 %, 0.00525 and 100.93 %, 0.06515 for NAP and ANT, respectively. Finally, a comparison between the proposed methods and high-performance liquid chromatography (HPLC) as a reference method was performed by analysis of variance (ANOVA). The results proved that there was no significant difference between them.
ISSN:0030-4026
1618-1336
DOI:10.1016/j.ijleo.2019.164010