OBJECTIVES: The proposed initiative presents the collaborative approach in classification of preliminary characteristics of diseases using sample clinical data that allows the integration of parallel processing in homogeneous grid-based environment. The research focuses on three objectives mainly:

1. To provide collaborative classification in homogeneous resources.
2. To conduct parallel processing in extraction of preliminary characteristics using electronic medical records (EMR) data.
3. To perform characterization for disease features in grid-based neural network classification.

METHODS: The study conducted on Globus (Grid) clustering network and interconnected with homogeneous resources as test bed. The integration of homogeneous sample diseases databases for execution of computational application were submitted to the GRAM service to the local scheduling system. The result for time and threads computation was computed on the test bed for homogeneous resources in grid platform with Feed-Forward Neural Networks. PRELIMINARY RESULTS: In the training phase, the diversity of clinical data features such as age, gender, race/ethnicity were imported as input to the Globus nodes with the aid of Globus automated scheduling for diseases’ characteristics classification. The coordination of resources aims to address the issue of optimization in distributed grid resources. The evaluation of outcome includes response time and co-allocation of multiple resources to meet complex clustering of diseases' characteristics using neural networks classification.