Constructing the Logical Regression Model to Predict the Target of Jianpi Jiedu Decoction in the Treatment of Hepatocellular Carcinoma

Figure 2

LASSO regression model was performed to further screen prognostic genes. (a) Lasso coefficient profiles with 20 characteristics. The coefficient profile was generated by comparing with the log (λ) sequence. 12 predictors of nonzero coefficients were obtained after selecting the optimal λ value (also known as the minimum λ value = 0.0324). (b) The selection of the optimal parameter (λ) in the LASSO model was cross-verified for 1000 times by the minimum standard. Draw a point vertical line at the optimal value by using 1 standard error (SE) of the minimum standard the minimum standard (1-SE standard). When choosing the optimal log (λ) value (the first vertical line), the model included 12 nonzero prediction values.

(a)

(b)