CUSTOM NAMED ENTITY RECOGNITION FROM CORPUS DATA USING CONDITIONAL RANDOM FIELD

Finance is a growing field which is concerned with the allocation of assets and liabilities over space and time. In any wealth management organization, several advise set documents are used where financial statements and client data are recorded. Therefore, the identification and recognition of custom named entities using Natural Language Processing could help the clients to understand and find relevant information from the extracted data. A new model using the method Conditional Random Field (CRF) is developed as it provides accurate results when compared to already existing model. The proposed model is fast, highly accurate, easy to install and use. This paper proposes a method to extract custom named entities from corpus data of financial domain using Conditional Random Field (CRF) and evaluates the effectiveness of the proposed method.