LFI ExS
The ExS package contains business intelligence elements and tools for building expert systems.
ID Cross Referencing
The ExS package includes a tool for automatically creating cross references between identifiers in different systems. Using our pattern matching engine we are able to find the best possible match based on the parameters you provide.The cross reference tool allows you to specify which attributes to check, the priority of each attribute, and the acceptable tolerances. The algorithm uses a multi-pass approach pulling out exact matches first before going on to find the best possible partial match.
Perceptron Classifier
A perceptron is a type of neural network used for simple data clustering. Ours is a single-layer implementation for linear classification which allows you to compartmentalize your data into regions based on the characteristics you provide. The classifier can them be used to determine what region(s) a subject case best fits into.Ranking Engine
A ranking engine is available which helps to sort opportunities, candidates, or other objects based on the criteria you supply.The engine supports two primary methods of ranking: simple and weighted. The simple approach utilizes a standard multi-column sort with each parameter having an order and ascending or descending option. The weighted approach calculates a ranking value for each object from the parameter data and supplied weightings. Objects are then sorted descending with the ones with the highest ranking value coming first.
Resource and Task Scheduler
The task scheduler automatically matches work that needs to be done with resources capable of doing it. The scheduler takes into account customizable prioritization so the most important tasks get assigned first.The scheduler also supports a flexible system of constraints which it can take into account when making assignments. Typical constraints include: can this resource actually perform this task, what is the earliest date on which this task can start, or when does this resource become available.
Route Optimizer
Our route optimizer is a unique solution to problem of planning the most efficient route. Unlike other solutions, we use a combination of genetic algorithms and independent path finders to dramatically improve the speed and accuracy of the results.A key feature of the solution is the support for path and site weighting. This allows us to take into account road conditions, preferential paths, and other external factors.
Self-Organizing Map
The ExS package includes a Self-Organizing Map (SOM) which is a type of neural network which uses dimensional reduction and vector quantization to generate a map of the data. Subject cases can be compared to the map to find where they best fit which can in turn drive recommendations, help find root causes, or make predictions about outcomes.The package includes not only the algorithms but also a number of visualization tools including a 3D bar chart, 2D confidence view, and the standard 2D map view.
The algorithm supports both a square and hex based neural map generation and display.
Statistical Pattern Matching
The Statistical Pattern Matching engine compares a subject case against a prepared knowledge base of known cases in order to find the best matches.The engine works by analyzing the knowledge base to determine means, standard deviations, and other properties for each of the attributes defined. Subject cases are then compared against each knowledge base case to determine how closely its variances match. The results of the analysis are returned ranked from best to worst match with information detailing which attributes lead to the decision made.