We continue the discussion on Text knowledge mining : An alternative to text data mining paper, We saw that the authors proposed the TKM as one based on abduction and includes techniques that generates new hypothesis. Data mining techniques are based on inductive inference. TDM can be both subjective and objective. When TDM is subjective, typically domain knowledge is involved. When TDM is objective, typically statistics is used. Together they represent simple facts in data structures from which new knowledge can be obtained. This is mining and even though there is strong association with inductive learning, the word induction is generally not used with reference to it  The prevalence of such approach implies that data mining has been useful. Data mining however is  different from text mining in that the former depends on data model while the latter depends on natural language. Data models have limited expressive power. Natural language is richer. It is able to represent not only simple facts but also general knowledge, from relation like rules to complex procedures. The disadvantages with natural language involve computationally expensive to manage structures, vagueness, ambiguity etc. There is no restriction to what kind of techniques whether deductive and/or adductive can be used with mining. By proposing text mining based on non-inductive inference, the authors propose a new direction. TKM is mining by reasoning using knowledge contained in text.TKM is a particular case of what the authors call knowledge mining. The latter defined as obtaining non-trivial previously unknown and potentially useful knowledge from knowledge repositories. The main difference between the definition of knowledge mining and the definition of data mining is that the former intends to discover knowledge from knowledge repositories while the latter attempts to discover knowledge from data repositories. The difference is the starting point, the basic material from which new knowledge is to be obtained. The starting point also makes the approach clear in that the former is deductive or adductive inference while the latter is inductive reference. Reading this it might feel like we are taking existing techniques and applying to a higher layer. In fact there can be other layers as well. Knowledge and data are not the only layers for example, the actions are based on knowledge. Having actions recommended from knowledge  is yet another strata that we haven't reached yet. Coming back to text mining, the phases of the TKM process are exactly the same as those of TDM i.e. first text refining for obtaining a computationally manageable intermediate form representing the text, a mining procedure to obtain new knowledge.and a final step for assessing and filtering the knowledge obtained.
#codingexercise
Double GetAlternateOddNumberRangeMin()(Double [] A)
{
if (A == null) return 0;
Return A.AlternateOddNumberRangeMin();
}