The canning of dika kernel (ogbono) soup into a ready‐to‐eat form has been recommended because of the special skills and time required in preparing this popular West African soup. This study investigated the thermal processing of ogbono soup, calculated the lethality (F) and process time, and predicted same using the artificial neural network (ANN) model. A dataset of the processing time, cold point, and retort temperature obtained during the heat penetration study was used for the model training, validating, and testing steps. There was a 54%, 71%, and 75% log‐reduction of the surviving Bacillus stearothermophilus spores after heat processing of the canned soup at 110, 116, and 121°C respectively. Heat penetration parameters revealed that canning the soup at 121°C for a total process time of about 31 min was the best thermal processing condition. Results also showed that the ANN model was a useful and high‐reliability tool in predicting the sterilization value of the canned soup, with a high degree of accuracy (R2 closer to 1) at all the canning temperatures considered in the study. Unsupervised learning of the neurons models to predict the output data from the input data was also suggested in the thermal canning processes.
Practical Applications
In the canning industry, the sterilization process is aimed to inactivate potentially harmful microorganisms, leaving the food product safe and of good quality. This research evaluated the canning parameters to process dika kernel (ogbono) soup. The total processing time of 31 min at 121°C obtained in the study, indicate that higher temperature and shorter time would produce a commercially sterile soup. However, the heat penetration test requiring higher temperature and longer processing times could necessitate increased energy costs. This research shows that the use of the machine learning (artificial neural network) model could accurately predict lethality rate (F‐value) better than the non‐linear model like the response surface methodology (RSM), providing an advantage of estimating process time, thus reducing operational costs.