A unique type of lymphoma that arises from germinal center B-cells is Hodgkin lymphoma (HL). This type of lymphoma differs from other cancer types by its exclusive features like immunohistochemical and morphological features. Most of the pathological labs retained highresolution images representing the tumors. The database having this information is upgrading day by day. Most of the prediction is made by gathering the one-dimensional data captured from the patients suffering from Hodgkin lymphoma. The proposed system utilizes an enhanced deep learning model for such gathered one-dimensional HL data to images. In this research, a deep learning model is designed for enhancing classification performance in terms of sensitivity and accuracy. The slides taken from HL tissue color images which is digitally recorded from pathological labs and the onedimensional data from the gene expression dataset are considered in this research. By using the hybrid JASNE algorithm (Jarvis’s Algorithm (JA) in addition with t- Distributed Stochastic Neighbor embedding(t-SNE) algorithm), the 1D image data is converted to high-resolution images. The classification process consists of three stages, two classification stages and one decomposition stage. Initially, fuzzy-based convolutional neural networks (FCNN) are used for fundamental classification. Then the molecular level of training is then performed by decomposing the dataset using Modified Empirical wavelet transform (MEWT) along with enhanced variational mode decomposition (EVMD). Then, using the hybrid multilayer perceptron-firefly algorithm (HMLP-FFA), another stage classification process is done. Compared to the other state-of-the-art methods, the proposed system provides better concert in terms of accuracy and sensitivity

Hodgkin Lymphoma, Pathological Labs, Jarvis algorithm, Multilayer Perceptron, Fuzzy based Convolutional neural networks