Bayesian network modelling on data from fine needle aspiration cytology examination for breast cancer diagnosis

The paper employed Bayesian network (BN) modelling approach to discover causal dependencies among different data features of Breast Cancer Wisconsin Dataset (BCWD) derived from openly sourced UCI repository. K2 learning algorithm and k-fold cross validation were used to construct and optimize BN structure. Compared to Naïve Bayes (NB), the obtained BN presented better performance for breast cancer diagnosis based on fine needle aspiration cytology (FNAC) examination. It also showed that, among the available features, bare nuclei most strongly influences diagnosis due to the highest strength of the influence (0.806), followed by uniformity of cell size, then normal nucleoli. The discovered causal dependencies among data features could provide clinicians to make an accurate decision for breast cancer diagnosis, especially when some features might be missing for specific patients. The approach can be potentially applied to other disease diagnosis. Introduction Breast cancer has become one of the most terrible cancer types in China. By 2008, there are more than 210 thounsand people died from breast cancer which has become the sixth killer in all kinds of cancer for Chinese women [1,2]. It is important for patients to be early diagnosed and treated using appropriate methods [3]. Bayesian network (BN) modelling approach has been extensively used for diagnosis of breast cancer. Kalet et al. [4] used a Bayesian model to detect misdiagnoses made at the initial stage of diseases, such as lung, brain and female breast cancer. Wang et al. [5] proposed a three-layer BN for the earlier diagnosis of breast cancer. Hassen et al. [6] used Bayesian network to estimate the risk of metastasis for breast cancer patients. However, there is a lack of report on quantitative analysis among different breast cancer features, neither any research founded about the causal dependency between any pair of different features. A clearly explanation of BN in a specific domain, such as healthcare, can well understand the disease pathology and provide valuable diagnostic decision for domain experts. In this paper, breast cancer features based on fine needle aspiration cytology (FNAC) derived from Breast Cancer Wisconsin Dataset (BCWD) of UCI repository [7] were analyzed using BN modelling approach to dicover the causal relationships 409 5th International Conference on Frontiers of Manufacturing Science and Measuring Technology (FMSMT 2017) Advances in Engineering Research (AER), volume 130