Neural Net Analytics for Managers and Analysts

Neural nets and related AI analytics of machine learning are an important emerging tool today. There are tools available today that provide for developing a neural net solution from the extract, transform and load stage through the model building stage and finally ending with the running and solution presentation stage. Many of the tool suites require data scientist professionals to develop the application. However, several tools have emerged that simplify the use of neural nets so a manger can more easily take advantage of their value without the intervention of a specialist.

Neural nets provide an alternative view of factors (called features in NN parlance) that influence some single objective value or grouping of values. With a neural net you get predictive results based on past factors.

The KCI preferred tool for easy neural net analysis, analytics and management support is the Predictor and Classifier tools available as part of the AI Trilogy suite from Ward Systems Group. These are developed, tested and simple neural nets that can be used for many of the analytical needs in an organization.

AI Trilogy is a basic neural net modeling management tool that provides managers and analysts with entry into the world of neural need application without the need to be a coder. The Predictor and Classifier nets use an excel spreadsheet as input. Most of the time the data is extracted from operational or other systems, cleaned and formatted for a neural net input and then used to train the net. Neural nets can show bias based on the data used to train them. Prediction and classification come from adding hypothetical feature values and observing the result generated based on the training set.

Like business analytics tools, neural net tools require focused workflows to get the most use out of them. KCI provides several workflows for key types of analysis typical of managers today.

Using AI Trilogy
AI Trilogy solutions can be applied from various perspectives, for example, at a high level (understanding the influence of external factors such as social trends on your strategies) or at a more detailed level (comparing a neural net analysis of process performance factors with a statistical approach).

Here are three of the solutions that most organizations seek on a regular basis. Each of these AI Trilogy insights are used to guide the user by providing the neural net steps and features needed by following easy to use workflows.

More in this category: « Decision Analytics for Managers and Analysts Strategic Alignment »
back to top
Process Performance Analysis


Which factor contributes the most to performance?

Flip to learn more
Process Performance Analysis
Processes have several performance features such as cycle time, queue time, transport time, efficiency, quality and so on.

Neural nets can identify which features have the greatest influence on efficiency.

Removing a feature from the analytic can show which secondary feature has the greatest impact and may be a better choice for improvement.

Prediction – Managers can vary the features, predict efficiency and find out what feature to focus on for best improvement

KPI Analysis


Which indicators are most important?

Flip to learn more…
KPI Analysis
Senior management often defines and provides KPI targets for operating units, product, locations and so on.

KPIs are then cascaded down to more detailed organization units, products.

Neural nets can identify which features related to the KPI have the greatest impact.

Neural nets also quantify the degree of that impact through correlation results.

Prediction - Managers can use the resulting net to predict KPI performance by varying the features thus focusing an improvement effort.

Hiring Classification


Keeping bias out of hiring

Flip to learn more…
Hiring Classification
Identifying well qualified candidates for different positions in an organization may have a variety of results.

To improve the selection process, features such as experience, degree type, age, marital status, current salary and so on are identified that reflect the needs of a position.

These features can then be used by a neural net to suggest whether a candidate should be hired, rejected, held for future consideration and so on.

When confronted with a large number of candidates, such analysis by a neural net is helpful in selecting valuable employees.