Abstract: A Hopfield neural network transfers information with feed-back connections. These are similar to magnetic materials where stability of the bit storage plays a crucial role in exchanging strength through the spin(+1 or -1) orientations. More stability helps in storage of multilevel data such as image data. A Complex Valued Hopfield Neural Network (CHNN) with a multi-stable Hopfield model has low stability in two-dimensional phase. Rotor Hopfield Neural Network (RHNN) added to CHNN increases its stability in multidimensional phase. Hyperbolic Hopfield Neural Network (HHNN) is an extension of CHNN by Clifford algebra. In our proposing system, we are extending the theories of stability between HHNN and RHNN by investigating this process through the projection rule. HHNN is independent of the resolution factor and there is a gradual increase in the noise tolerance. Thus, it is comparatively more stable than RHNN.

Keywords: CHNN, RHNN, HHNN, Projection Rule

| DOI: 10.17148/IARJSET.2019.6707