Neural Networks were developed to identify different phases in the growth cycle of a recombinant strain of Bacillus subtilis. The network performance was enhanced by adding the first and second derivatives of the cultivation sample absorbance. The new architecture was then coupled to a Jordan network with locally recurrent processing elements in order to forecast the stages of the fermentation based on data collected during the first 5 hours.
All Science Journal Classification (ASJC) codes
- Applied Microbiology and Biotechnology