Comparison between Artificial Neural Network and Adaptive Neuro-Fuzzy Inference System For The Baryon-to-Meson Ratios in Proton-Proton Collisions

TL;DR

This paper compares the effectiveness of ANFIS and ANN in predicting baryon-to-meson ratios in high-energy proton-proton collisions, demonstrating that ANFIS outperforms ANN and theoretical models in accuracy and training speed.

Contribution

It introduces a comparative analysis of ANFIS and ANN for particle ratio prediction, highlighting ANFIS's superior performance and faster training in high-energy collision data modeling.

Findings

ANFIS outperforms ANN and theoretical models in prediction accuracy.

Both ANFIS and ANN match experimental data well.

ANFIS trains faster than ANN and other models.

Abstract

This article presents two systems that can simulate and predict Particles ratios created in high energy proton-proton (pp) collisions as a function of transverse momentum and the center-of-mass energy. An adaptive neurofuzzy inference system (ANFIS) and an artificial neural network (ANN) system are the systems in discussion. The ANFIS and ANN simulation results for training particles ratios as evaluated with training data points revealed an excellent match to the experimental data. The ANFIS and ANN's prediction abilities were also tested using data points that were not included in training and they performed well. The results clearly show that these methods are capable of extracting collision information and that they are helpful. Also, ANFIS and ANN results were compared with additional theoretical results (PYTHIA (CR Mode), HERWIG7, PYTHIA, PYTHIA8 (Monash), and EPOS-LHC). It is found that the ANFIS shows better performance and is trained more quickly than the ANN system and other theoretical models.