Mohamed HAFIDI ^1* Loubna El Fels ¹ Ahmed Naylo ¹ Martin Jemo ² Nidal Zrikam ¹ Ali Boularbah ¹ Yedir Ouhdouch ¹

• ¹ Cadi Ayyad University, Marrakech, Morocco
• ² Mohammed VI Polytechnic University, Ben Guerir, Morocco

We explored the enzymatic activities produced by living microorganisms of lignocellulosic compost mixtures during14 weeks. The phosphatase activity was higher for Mix-A (398.7 µg PNP g -1 h -1 ) than Mix-B (265.3 µg PNP g -1 h -1 ), while Mix-B (103.3 µg TPF g -1 d -1 ) exhibited greater dehydrogenase content than Mix-A (72.3 µg TPF g -1 d - 1 ). The gradual increase in the dehydrogenase from the compost Mix-A implies that high lignocellulosic substrate requires gradual buildup of dehydrogenase activity to turn the waste into mature compost. A higher pick of urease with a maximum activity of 151.5 and 122.4 µg NH4-N g -1 h -1 were reported, respectively for Mix-A and B. The machine learning well predicted the compost quality based on NH3/NO3, C/N ratio, decomposition rate and, humification index (HI). The root mean square error (RMSE) values were 1.98, 1.95, 4.61%, and 4.1 for NH + 3/NO -3, C/N ratio, decomposition rate, and HI, respectively.The coefficient of determination between observed and predicted values were 0.87, 0.93, 0.89, and 0.94, for the r NH3/NO3, C/N ratio, decomposition rate, and HI. Urease activity significantly predicted the C/N ratio and HI only. The profile of enzymatic activity is tightly linked to the physico-chemical properties, proportion of lignocellulosiccomposted substrates.