How do you train the Neural Credit Assistant?
The system is not programmed with any preexisting rules or structure; it actually learns through experience and trial and error. The neural system uses a training procedure through which it is repeatedly shown examples of how the various rating agencies have rated the senior debt of companies and the financial ratios associated with those companies. The model will start out assigning weights to each of the inputs into the model and for each example calculate how closely the resulting rating was to the desired rating and then sum these differences to arrive at a score for this set of weightings. This process will be repeated interactively to adjust the weightings to minimize the errors of the model. Based on these examples, the system "learns" the nature of the relationship between the financial ratios and the ratings that have been assigned. Each month the system is retrained by including not only the examples that were available to the system last month, but also the new examples available from companies releasing new financial data, new companies being rated, and companies being upgraded and downgraded. This allows the system to continuously learn and improve.

How can I review a company not rated by the rating agencies?
One of the benefits of the NCA's credit classification ability is that it provides an objective tool to analyze companies regardless of whether or not a major rating agency has rated the company. With the NCA investors needing to compare the credit quality of companies no longer need to limit their analysis to publicly rated companies. We currently maintain a database of an additional 7,000 companies that we can provide comparable NCA rating for review. We do offer an additional product that provides investors with the ability to rate companies not currently covered. If you have an interest in a specific company not currently reported in the Neural Credit Assistant or the rating product please contract me at the e-mail or phone number below.

More Frequently Asked Questions

FAQ Page: 1 2 3 4 5