Gradual-increase extraction of target baskets as preprocess for visualizing simplified scenario maps by KeyGraph

Abstract

KeyGraph is one of the powerful methods to support mining some knowledge from huge dataset because of its visualization mechanism. It presents the dataset in network diagram with representative events and their relationships. The data analyst reads its relationships, and supposes scenarios from them. This is conceptually very simple process, but it becomes more difficult when the diagram becomes complex. In this paper, to overcome this difficulty, we develop the pre-process to generate simple network KeyGraph and the scenario supposing process which repeats our pre-process, generation of KeyGraph and supposing scenarios.Keywords Data mining · KeyGraph · Scenario · Data preprocessing · Gradual-increase extraction of baskets Background and objectivesData mining is the method to obtain knowledge from huge dataset. Its most important element is how useful, understandable, and new knowledge we can obtain (Fayyad et al. 1995). In case which the dataset uses small number of variables, we can obtain knowledge relatively

Qualitative Chance Discovery – Extracting competitive advantages

“…A soft risk map [4] of natural disaster is a map aimed at the visualization of natural disaster's risk level. KeyGraph [15] has been used to represent the visualized data structure in Chance Discovery, but the data structure is often too complex to understand [12]. In other words, how to develop a process, which is not only capable of sifting out relevant but low frequency data, but also capable of combining visualizing technique with Grounded theory or qualitative meta-synthesis, to help users in the effort of searching future chances, has become an important research topic.…”

“…Depending on bridge and connecting clusters, users can explain the chance scenario [14]. But in reality, the data structure may be too complex to understand [12]. Meanwhile, the qualitative analysis of Grounded theory is a useful method for creating conceptual scenario.…”

Identifying Consensus Tags in Social Tagging Systems

“…KeyGraph represents the relation between words as a relation graph, with which we could extract keywords. The basic algorithm of KeyGraph is [14]: a) Generating basket-word array. Put basket number in the firstrow, and put all words used in all of the baskets in the first column.…”

RFID Tags as Technology for Value Sensing in Real Space Market

“…Here, a scenario is a series of actions and events which occur under a coherent 528 context. As in [2,3,4], a scenario map is useful for aiding chance discovery i.e., to detect an event significant for decision making as the picking of 20$ cheese. This effect works more finely, if the customer's behavior to pick an item out of a shelf can be taken in the data.…”