Classification of cancer patients based on elemental contents of serums using bidirectional associative memory networks

Bidirectional associative memory (BAM) was applied for diagnosing cancer ba sed on the elemental contents in serum samples. The serum samples were take n from clinical hospitals in north-east region of PR China and the elementa l contents in serum samples were analyzed by inductively coupled plasma ato mic emission spectrometry (ICP-AES). The elemental contents of the sample w ere encoded to bipolar input values for BAM computation. The BAM method was verified with independent prediction samples by using the 'cross-validatio n' method. The networks can be used to discriminate of all cancer patients from non-cancer patients at rate of 100%. A comparison study using BAM and multi-layer feed-forward neural network was made, better results were obtai ned using BAM networks. The effects of threshold values and output nodes of the BAM network were investigated and related problems were discussed. Res ults showed that the BAM would be applied to elemental analysis of serums a nd be promising method for diagnosis of cancer. (C) 2001 Elsevier Science B .V. All rights reserved.