**2.3. Assisted query formulation**

Due to the big data contingency and complex schematic structure of data, sensible formalisms of query is required for complex information retrieval which is mastered by a small group of users usually. Most users in real life apply brute force approaches which manipulate data by hand as they have little knowledge regarding query formulation. Assisted query formulation techniques are proposed to resolve these issues. These techniques assist the user by suggesting some query terms for subsequent formulation of incremental queries and reduction of irrelevant data retrieval. Fundamental operations such as equijoin and semijoin [11] are characterized for the formation of Boolean membership queries in polynomial time. A user membership driven learning algorithms [2] can also serves better formulation for simple Boolean queries. Many other formulation techniques for query construction such as locate minimal project join queries, discovering query approach [34] etc. are developed to answer query formulation similar to example tuples [27].

from *Q* to create variations *Qi*

**3.1. Proposed approach**

tion. Initially query Qi

specifically, *S* = *D1 × D2 × …, × Dd*

**Figure 3.** A conceptual example of query morphing.

query Qi

and retrieved outcomes analogous to the user history log.

. A conceptual example is shown in **Figure 3**. The result set

Query Morphing: A Proximity-Based Approach for Data Exploration

http://dx.doi.org/10.5772/intechopen.77073

153

returned by user's original query request is painted by the small red circle. Possible additional relevant result sets of user's interest for queries are explored in the large data spectrum, providing that result set belongs to surrounding closed region of the original request. Orange elliptic represents the query results that correspond to variations of original query request. After analyzing results user may formulate another query to shift interest towards another query result as shown in the right portion of **Figure 3**. A new region of query result of the user's request includes both new and previous variations of the query. A data space expedition and feedback incorporation is observed for the query reformulation and additional relevant data suggestion in this approach. The additional relevant data objects are retrieved by performing exploration and exploitation in available database. The properties of traditional techniques are also incorporated in our query morphing approach as user's search request

Creating transformation of input like text, image, data etc. is a fundamental process in computer science called morphing [10]. We analogous our query with the morphing inputs and named our reformulation algorithm as 'Query morphing' [41]. Our algorithm helps user in formulation of intermediate queries by creating variants/transformation of the original search query. The assistance to the user will be based on the optimal query reformulations derived during exploration and exploitation of dataspace. The proposed algorithms are developed by considering 'Query-Result- Review-Query' paradigm of computing [7, 15, 39]. The design

Our query reformulation approach can be seen into two sub activities, one is tradition query processing the other one is generation of morphs that derive intermediate query reformula-

query processing mechanism by the DBMS. Data objects retrieved after processing initial

non-overlapping rectangular cells and exploited for subsequent interactions. If we say more

are identified on *d*-dimensional space that is already created and partitioned into

will be validated and processed by the query engine in the traditional

be a *d*-dimensional space where *D* = {*D1*

*, D2*

*,…, Dd*

} be a

framework for the same is conveyed in following section and shown in **Figure 4**.

We termed our approach as 'Query morphing' because in literature a traditional method, morphing points transformation of inputs e.g. Data Morphing [20], Image morphing [10, 24]. Similarly, a small transformation of user queries are also carries out in our approach. We realized that the success of our approach is mainly rely on effective database exploration and user participation. The properties of traditional techniques are also incorporated in our query morphing approach as user's search request and retrieved outcomes analogous to the user history log.
