Fig. 2 | Scientific Reports

Fig. 2

From: Development and validation of machine learning models for diagnosis and prognosis of lung adenocarcinoma, and immune infiltration analysis

Fig. 2

Feature selection. (A) Feature selection of mRNAs using RF. RF.analysis classifies the levels of importance of mRNAs. The X‐axis represents RF mean decrease in accuracy and gini. The Y‐axis represents a ranking of variables was obtained from RF on the basis of mean decrease in accuracy and gini coefficients. (B) Performance (error rate) per number of tree generated by the RF algorithm. The x-axis shows the number of trees, Y-axis is the error rate given by RF (out of bag error estimation from 800 trees). Black, red and green lines correspond to the gross distribution, stage I LUAD distribution and adjacent normal tissue distribution, respectively. (C) Lasso analysis results of mRNAs. The lower horizontal axis represents lambda value, and the upper horizontal axis scale represents the number of variables in the lasso model, the regression coefficient (x) of which is not 0. (D) The trajectory of each independent variable, the horizontal axis represents the log value of the independent variable lambda, and the vertical axis represents the coefficient of the independent variable. (E) A venn diagram of two feature selection methods. (F) Feature selection of miRNAs using RF. RF analysis classifies the levels of importance of miRNAs. (G) Performance (error rate) per number of tree generated by the RF algorithm. (H) Lasso analysis results of miRNAs. (I) The trajectory of each independent variable, the horizontal axis represents the log value of the independent variable lambda, and the vertical axis represents the coefficient of the independent variable. (J) A venn diagram of three feature selection methods. (K) Feature selection of lncRNAs using RF. RF analysis classifies the levels of importance of lncRNAs. (L) Performance (error rate) per number of tree generated by the RF algorithm. (M,N) Two-dimensional projections using four of the SVM features. The support vectors and non-support vectors are denoted with triangles and circles, respectively. Red areas represent predicted positive regions, yellow areas represent predicted negative regions. (O) A venn diagram of two feature selection methods.

Back to article page