Doreswamy, Hemanth K S and Manohar M G, Mangalore University, India

Nowadays numerous interestingness measures have been proposed to disclose the relationships of attributes in engineering materials database. However, it is still not clear when a measure is truly elective in large data sets. So there is a need for a logically simple, systematic and scientific method or mathematical tool to guide designers in selecting proper materials while designing the new materials. In this paper, linear regression model is being proposed for measuring correlated data and predicating the continues attribute values from the large materials database. This method helps to find the relationships between two sub properties of mechanical property of different types of materials and helps to predict the properties of unknown materials. The method presenting here effectively satisfies for engineering materials database, and shows the knowledge discovery from large volume of materials database. Studying on regression analysis suggests that data mining techniques can contribute to the investigation on materials informatics, and for discovering the knowledge in the materials database, which make the manufacturing industries to hoard the waste of sampling the newly materials.

Mechanical property, Materials database, Knowledge discovery , Regression ,Correlation .