Build and help me to determine the technical requirements and most optimizing features for a unique semantic networking computer program

I need someone to build a computer program for a project I am working on and also suggest the most optimizing features for it considering its goal. By my knowledge the closest thing I can compare it to within the arena of computer science is a semantic network with some added features. I need a program that can represent and organize large amounts data or knowledge in a way that prioritizes and highlights the coherence of the system. This system has a quadrilateral or polygonal framework with 18 spaces for concepts or abstractions that will be evenly arranged around its border. Each concept will have a single word which will serve as the essential theme for that conceptual bubble and the fundamental avenue by which the program is to compile all related words and synonyms together (from a number of different thesauri/lexical databases). This compiling will be extended to the synonyms of its synonyms and the related words of its related words several layers down, to form something of a semantic reference field for each word entered. The reference field for each of these concepts will have a collection of related terms connected to whatever primary word originally entered. Each of these concepts within the framework will be interconnected to others that are in some way symmetrically aligned with it, the program is to identify and label the kind of semantic relations formed in the framework which will always be the same for the same type of symmetrical alignments and highlight the particular instances or examples of that kind of relation found between words located in each of the concepts semantic reference field. The way a set of concepts relate to an individual one, should be used by the software to contextualize and compartmentalize the compiled words within that individual concepts semantic reference field and to highlight the most relative elements within it. Through these relations the program should be able to bilaterally infer things about each concept involved that supports or corroborates particular words within each concepts field of reference, which the program would then display in some way showing it to be a greater pertinence. The summary and general aim of the program is to create concepts through entered keywords, determine and highlight every type of semantic relation between each of these concepts and to use the way each of these concepts specifically connect in word to word relations to characterize the particular facets of those concepts (amongst its compiled semantics reference field). The ultimate aim of all this is to find and highlight the connections that allow the easiest comprehension of the coherence of the system or to put in another way the greatest continuity within the fragmented coherence of the system. There are a few other features it would need in the area of search or suggesting a word concept through inferred data but this is the general idea.