Algorithm 1 Decision tree generation algorithm

Construct training set from sensor data $X$ is a matrix of $R\times M$, $X_{ij}$ represents the j-th feature of the i-th sample. $Y$ is a matrix of $R\times 1$, $Y_i$ denotes the class label of the i-th sample. Build a decision tree    If all the sample values of $X$ are the same, or all the class labels of $Y$ are the same, or $R<2$, a leaf node is generated, and the class of this node is the class of the most number in $X$.    else:     Select $m$ randomly from $M$ features.     Among these $m$ features, the maximum information gain is denoted as $p$.     If: the value of feature $p$ is discontinuous      $V$ is any value of $p$      $X_V$ is used to represent the sample whose feature $p$ takes $V$, $Y_V$ is the corresponding class.      $Childv=Generate\left(X_V,Y_V\right)$      Return a decision tree node     If: the value of feature $p$ is continuous      t is the best split threshold.      If: $X_{LO}$ represents a sample set whose values of feature $p$ is less than $t$, and $Y_{LO}$ is its corresponding class.      $Childlo=Generate\left(X_{LO},Y_{LO}\right)$      If: $X_{HI}$ represents a sample set whose values of feature $p$ is greater than or equal to $t$, and $Y_{HI}$ is its corresponding class.        $Childlo=Generate\left(X_{HI},Y_{HI}\right)$      Return a decision tree node