Rajesh Balakrishnan

Flowchart knowledge extraction on image processing

In the paper, we present an approach of image processing analysis to extract flowchart information from digital imagery. Firstly, flowchart imagery is processed to extract the text components and then extract the geometrical shapes components. We analyze text, and various geometrical shapes present in flowchart and carry out a variety of processes such as image segmentation, shape description, text and geometric components extraction, recognition and linking. The text components are extracted and then geometrical components are extracted. we also proposed a auto directional transformation of contour chain method for shape description. The internal relationship between the components is set up by tracing the flow lines which connect different components. Thus a flowchart is correctly extracted. The extracted components are output to metadata (XML format) which is machine readable. These metadata can be archived, store as knowledge base or shared with others. Finally, an example is presented and the results show that the proposed technique is efficient.

Semantic wiki aided business process specification

This paper formulates a collaborative system for modeling business application. The system uses a Semantic Wiki to enable collaboration between the various stakeholders involved in the design of the system and translates the captured intelligence into business models which are used for designing a business system.

Leveraging Call Center Logs for Customer Behavior Prediction

Most major businesses use business process outsourcing for performing a process or a part of a process including financial services like mortgage processing, loan origination, finance and accounting and transaction processing. Call centers are used for the purpose of receiving and transmitting a large volume of requests through outbound and inbound calls to customers on behalf of a business. In this paper we deal specifically with the call centers notes from banks. Banks as financial institutions provide loans to non-financial businesses and individuals. Their call centers act as the nuclei of their client service operations and log the transactions between the customer and the bank. This crucial conversation or information can be exploited for predicting a customers behavior which will in turn help these businesses to decide on the next action to be taken. Thus the banks save considerable time and effort in tracking delinquent customers to ensure minimum subsequent defaulters. Majority of the time the call center notes are very concise and brief and often the notes are misspelled and use many domain specific acronyms. In this paper we introduce a novel domain specific spelling correction algorithm which corrects the misspelled words in the call center logs to meaningful ones. We also discuss a procedure that builds the behavioral history sequences for the customers by categorizing the logs into one of the predefined behavioral states. We then describe a pattern based predictive algorithm that uses temporal behavioral patterns mined from these sequences to predict the customers next behavioral state.