Erythemato-squamous diseases are a general designation of six common skin diseases, of which the differential diagnosis is a difficult problem in dermatology. This paper presents a new method based on virtual coding for qualitative variables and multinomial logistic regression penalized via elastic net. Considering the attributes of variables, a virtual coding is applied and contributes to avoid the irrationality of calculating nominal values directly. Multinomial logistic regression model penalized via elastic net is thence used to fit the correlation between the features and classification of diseases. At last, parameter estimations can be attained through coordinate descent. This method reached accuracy rate of 98.34%±0.0027% using 10-fold cross validation in the experiments. Our method attained equivalent accuracy rate compared to the results of other methods, but steps are simpler and stability is higher.
Signal classification is a key of brain-computer interface (BCI). In this paper, we present a new method for classifying the electroencephalogram (EEG) signals of which the features are heterogeneous. This method is called wrapped elastic net feature selection and classification. Firstly, we used the joint application of time-domain statistic, power spectral density (PSD), common spatial pattern (CSP) and autoregressive (AR) model to extract high-dimensional fused features of the preprocessed EEG signals. Then we used the wrapped method for feature selection. We fitted the logistic regression model penalized with elastic net on the training data, and obtained the parameter estimation by coordinate descent method. Then we selected best feature subset by using 10-fold cross-validation. Finally, we classified the test sample using the trained model. Data used in the experiment were the EEG data from international BCI Competition Ⅳ. The results showed that the method proposed was suitable for fused feature selection with high-dimension. For identifying EEG signals, it is more effective and faster, and can single out a more relevant subset to obtain a relatively simple model. The average test accuracy reached 81.78%.