Multi-graph-Based Intent Hierarchy Generation to Determine Action Sequence

All actions are result of scenarios. While we analyze different information pieces and their relationships, we notice that actions drive intents and intents drive actions. These relationships among intents and actions can decode the reasoning behind action selection. In the similar way, in stories, news and even email exchanges these relationships are evident. These relationships can allow us to formulate a sequence of mails irrespective of change in subject and matter. The continuity among concepts, drift of concepts, and then reestablishing of original concepts in conversation, email exchanges, or even series of events is necessary to establish relationships among various artifacts. This paper proposes a multi-edge fuzzy graph-based adversarial concept mapping to resolve this issue. Instead of peaks, this technique tries to map different valleys in concept flow. The transition index between prominent valleys helps to decide your order of that particular leg. The paper further proposes a technique to minimize this using fuzzy graph. The fuzziness in graph represents the fuzzy relationships among concepts. The association among multiple such graphs helps to represent the overall concept flow. The dynamic concept flow represents the overall concept flow across the documents. This algorithm and representation can be very useful to represent lengthy documents with multiple references. This approach can be used to solve many real-life problems like news compilation, legal case relevance detection, and associating assorted news from the period under purview.