What is Visual and Audio Data Mining?

Visual data mining finds implicit and beneficial knowledge from huge data sets using data and knowledge visualization methods. The human visual system is managed by the eyes and brain, the latter of which can be think of as a dynamic, largely parallel processing and reasoning engine including a huge knowledge base.

Visual data mining can be considered as a unification of two disciplines such as data visualization and data mining. It can also associated with computer graphics, multimedia systems, human computer interaction, pattern identification, and highperformance computing.

In general, data visualization and data mining can be integrated in the following ways −

Data visualization − Data in a database or data warehouse can be considered at multiple levels of granularity or abstraction, or as several combinations of attributes or dimensions. Data can be displayed in several visual forms, including boxplots, 3-D cubes, data distribution charts, curves, surfaces, link graphs, etc.

Data mining result visualization − Visualization of data mining results is the presentation of the results or knowledge acquired from data mining in visual forms. Such forms can involve scatter plots and boxplots (acquired from descriptive data mining), and decision trees, association rules, clusters, outliers, generalized rules, etc.

Data mining process visualization − This kind of visualization presents the multiple processes of data mining in visual forms so that users can view how the data are derived and from which database or data warehouse they are extracted and how the selected data are cleaned, integrated, preprocessed, and mined. Furthermore, it can also show which approach is selected for data mining, where the results are saved, and how they can be considered.

Interactive visual data mining − In interactive visual data mining, visualization tools can be used in the data mining process to provide users create intelligent data mining decisions. For instance, the data distribution in a group of attributes can be showed using colored sectors (where the whole space is defined by a circle). This display supports users decide which sector must first be selected for classification and where the best split point for this sector can be.

Audio data mining need audio signals to denote the patterns of data or the features of data mining outcomes. Although visual data mining can disclose interesting patterns utilizing graphical displays, it needs users to concentrate on watching patterns and recognizing interesting or novel characteristics inside them.

If patterns can be changed into sound and music, rather than watching pictures, it can listen to pitches, rhythms, tune, and melody to recognize anything interesting or unusual. This can relieve various burden of visual concentration and be more comfortable than visual mining. Hence, audio data mining is an interesting counterpart to visual mining.