Structuring of Unstructured Data from Heterogeneous Sources

Objectives: To develop a new data gathering processing under Big Data Perspectives. convert unstructured text into structured format by not missing out any available. Methods: The is preprocessed using modified stemming and tokenization. From the output, proposed Term Frequency-Inverse Document Frequency (TF-IDF) N-gram features are derived. Unstructured considered from multiple sources like twitter, consumer complaints news blog. Findings: model with extant TF-IDF has exposed relatively high Mean Average Error (MAE) value which 1.4325 when compared to without optimization be 0.5197. Novelty: novelty of research work process where dictionary checking added improved feature extraction, interclass dispersion coefficient computed in features. Keywords: Natural language processing; Structured data; Feature extraction