RESEARCH ON BP NEURAL NETWORK APPLIED TO PREDICT COTTON FABRIC HANDLE

摘要

Their handle can assess inner property of fabrics. The subjective assessment is that the fabric is ranked or definitely described after its handle is assessed by sense and graded using subjective judgement, which consist of two methods, i. e. absolute judgement and ranking judgement. The objective assessment is that fabric is assessed according to eighteen fabric physical indexes of KES-FB. This paper mainly discusses the relation between subjective assessment and objective assessment by constructing the predictable model. The fabric handle can be predicted by artificial neural network model. Applying this method, the optimal power index and deviation can be got without constructing the mathematics model. In addition, this method can exactly predicts the fabric handle because the results are good enough. The artificial neural network predicting model can not only provides the exact index for fabric produce and research but also assess the fabric handle exactly and guide the fabric design.