Wyniki wyszukiwania - BazTech

1

Improving Diesel Engine Reliability Using an Optimal Prognostic Model to Predict Diesel Engine Emissions and Performance Using Pure Diesel and Hydrogenated Vegetable Oil

Žvirblis Tadas, Hunicz Jacek, Matijošius Jonas, Rimkus Alfredas, Kilikevičius Artūras, Gęca Michał

Eksploatacja i Niezawodność

|

2023

|

Vol. 25, no. 4

art. no. 174358

EN

The reliability of internal combustion engines becomes an important aspect when traditional fuels with biofuels. Therefore, the development of prognostic models becomes very important for evaluating and predicting the replacement of traditional fuels with biofuels in internal combustion engines. The models have been made to model AVL 5402 engine emission, vibration, and sound pressure parameters using a three-stage statistical regression models. The fifteen parameters might be accurately predicted by a single statistic presented here. Both fuel type (diesel fuel and HVO) and engine parameters that can be adjusted were considered, since this analysis followed the symmetry of the methods. The data analysis process included three distinct steps and symmetric statistical regression testing was performed. The algorithm examined the effectiveness of various engine settings. Finally, the optimal fixed engine parameter and the optimal statistic were used to construct an ANCOVA model. The ANCOVA model improved the accuracy of prediction for all fifteen missing parameters.

2

Cat swarm optimization algorithm tuned multilayer perceptron for stock price prediction

Chandar S. Kumar, Punjabi Hitesh

Computer Assisted Methods in Engineering and Science

|

2022

|

Vol. 29, no. 1-2 spec.

145--160

EN

Due to the nonlinear and dynamic nature of stock data, prediction is one of the mostchallenging tasks in the financial market. Nowadays, soft and bio-inspired computing algorithms are used to forecast the stock price. This article assesses the efficiency of thehybrid stock prediction model using the multilayer perceptron (MLP) and cat swarm optimization (CSO) algorithm. The CSO algorithm is a bio-inspired algorithm inspired bythe behavior traits of cats. CSO is employed to find the appropriate value of MLP parameters. Technical indicators calculated from historical data are used as input variablesfor the proposed model. The model’s performance is validated using historical data notused for training. The model’s prediction efficiency is evaluated in terms of MSE, MAPE, RMSE and MAE. The model’s results are compared with other models optimized byvarious bio-inspired algorithms presented in the literature to prove its efficiency. The empirical findings confirm that the proposed CSO-MLP prediction model provides the bestperformance compared to other models taken for analysis.

3

A Probabilistic-Fuzzy System Applied To Modeling Of Time Series

Błaszczyk K.

Zeszyty Naukowe. Elektryka / Politechnika Opolska

|

2009

|

z. 62

15-16

EN

The paper describes an approach to time series modeling using probabilistic-fuzzy system. Results of prediction for experimental data are compared with the results obtained from fuzzy inference system FIS and neuro-fuzzy system ANFIS.