Today we continue discussing TKM techniques. Let us review the discovery of contradictions. Let us say res(ti) is a procedure performing resolution on the set of clauses obtained from ti, giving the value  false in case it finds a contradiction. Let ti, tj be the concatenation of texts ( the corresponding clauses ) ti and tj. Let PT be the set of all possible concatenations of subsets of T of any size. V represents the set of concatenations that do not contain contradictions.  Then the algorithm for discovering contradictions is as follows. We find a candidate V1 with text that has no contradictions. We concatenate it to V. This is our initialized set. For the rest of the n iterations, we find text t'  candidate and t'' initialized such that they are from disjoint sets and their combination doesn't contain a contradiction. We check with all the subsets of text containing the candidate to rule out contradictions and perform a union for non-contradiction. Finally, with the set of all possible concatenations of subsets of T of any size and with V that contains a set without contradictions, we can exclude the ones that do contain contradictions.
We now look at the challenges of Text knowledge mining. There is a good analogy with data mining in that existing techniques worked for data mining  and similarly there are existing techniques that can work for TKM. These include the ones for knowledge representation, reasoning algorithms for performing deductive and abductive inference, and knowledge based systems, as well as natural language processing techniques. As in the case of data mining, these techniques must be adapted for TKM.  In addition, new techniques may be needed.  We look at some of these challenges as well.
Knowledge based systems and TKM systems have very different objectives that affect the way techniques coming from knowledge representation and reasoning can be adapted to TKM. In a knowledge based system, a knowledge base is built containing the knowledge needed by the system to solve a specific problem. In TKM, the knowledge managed by the systems is collected from texts each of which is treated as a knowledge base as in historical or normative reports as opposed to something that builds a knowledge system. Another difference is that the knowledge based systems are intended to give answer to every possible question, or to solve any possible problem posed to the system while in TKM, new hypothesis and potentially useful knowledge is derived from a collection of text.  In fact, TKM may exclude or not know about a set of knowledge pieces.
#codingexercise
Double GetAlternateOddNumberRangeMode()(Double [] A)

{

if (A == null) return 0;

Return A.AlternateOddNumberRangeMode();

}