Study of a screening system for mild cognitive impairment based on machine learning model / 上海交通大学学报（医学版）

Objective:To evaluate the reliability and validity of a computerized cognitive assessment system designed for screening mild cognitive impairment (MCI), and compare the screening accuracy among constructed different machine learning classification models. Methods:A group of random stratified samples of over 55 years old residents in the communities, nursing homes and memory-clinics from Shanghai and Henan were selected to assess their cognitive status using Montreal Cognitive Assessment (MoCA) by well-trained investigators. The reliability and validity were assessed by intrinsic consistency analysis and factor analysis, respectively. Taking the results of MoCA as standards, four machine learning classification algorithms, i.e., naïve Bayesian classification model, random forest classifier, Logistic regression classifier, and K-nearest neighbor classifier, were compared in accuracy and area under curve (AUC). Results:A total of 359 participants were included, the median age of whom was 63 years old. And 82.80% of them were secondary school graduates or below. According to the results of MoCA, 147 of them might be MCI. The Cronbach's α and KMO of this system were 0.84 and 0.78, respectively; Bartlett's sphericity test was significant (P<0.05); thirteen common factors could explain 75.10% of the system. The best classification model was naïve Bayesian classification model, and its accuracy and AUC were 88.05% and 0.941, respectively. Conclusion:The new designed computerized cognitive assessment system has been proved to be reliable and valid. The naïve Bayesian classification model has good classification accuracy.

Classification Algorithm Performance Study on Diabetes Electronic Medical Records / 医学信息学杂志

The paper preprocesses the data including basic information,admission and discharge record and progress note of diabetes Electronic Medical Records (EMR),implementing decision tree,Artificial Neural Network (ANN),Naive bayesian and K-Nearest Neighbor (KNN) classifications respectively on data that have been processed with Weka 3.9.The result shows that Naive bayesian classification,which is superior to the others in predicting and classifying such data,can provide basis for the classification and prediction of diabetes.

Tree-Augmented NaÏve Bayesian network model for predicting prostate cancer / 中华男科学杂志

<p><b>Objective</b>To evaluate the integrated performance of age, serum PSA, and transrectal ultrasound images in the prediction of prostate cancer using a Tree-Augmented NaÏve (TAN) Bayesian network model.</p><p><b>METHODS</b>We collected such data as age, serum PSA, transrectal ultrasound findings, and pathological diagnoses from 941 male patients who underwent prostate biopsy from January 2008 to September 2011. Using a TAN Bayesian network model, we analyzed the data for predicting prostate cancer, and compared them with the gold standards of pathological diagnosis.</p><p><b>RESULTS</b>The accuracy, sensitivity, specificity, positive prediction rate, and negative prediction rate of the TAN Bayesian network model were 85.11%, 88.37%, 83.67%, 70.37%, and 94.25%, respectively.</p><p><b>CONCLUSIONS</b>Based on age, serum PSA, and transrectal ultrasound images, the TAN Bayesian network model has a high value for the prediction of prostate cancer, and can help improve the clinical screening and diagnosis of the disease.</p>