In the current work various types of epoxy composites were added to concrete to enhance its effectiveness as a gamma- ray shield. Four epoxy samples of (E/clay/B4C) S1, (E/Mag/B4C) S2, (EPIL) S3 and (Ep) S4 were used in a comparative study of gamma radiation attenuation properties of these shields that calculating using Mont Carlo code (MCNP-5). Adopting Win X-com software and Artificial Neural Network (ANN), µ/ρ revealed great compliance with MCNP-5. By applying (µ/ρ) output for gamma at different energies, HVL, TVL and MFP have been also estimated. ANN technique was simulated to estimate (µ/ρ) and dose rates. According to the results, µ/ρ of all epoxy samples scored higher than standard concrete. Both S2 and S3 samples having higher values of µ/ρ, show minimum dose rate values. (µ/ρ) and RPE% values were enhanced, the concrete containing E/Mag/B4C (S2) had the best results, while the concrete containing Ep (S4) provide the worst results. The ANN prediction results take 15 sec for estimating gamma doses corresponding to seventeen shield thicknesses, while the theoretical MCNP-5 results took approximately between 7 to 10 hours for five gamma doses. ANN provides excellent predictions with a high degree of correlation depending on increasing the number of attenuation parameters used in the training process. Also, it predicts gamma dose rates for a large number of shield thicknesses that cannot be calculated theoretically in a very short time. This supports, the created epoxy composite offers good attenuation properties for many shielding applications and could be proposed as an injecting mortar for cracks in biological shields and the walls of diagnostic and radiotherapy rooms. However, further investigations are planned for different filler ratios, for comparison purposes, in order to reach optimal shielding properties
