Merge pull request #1 from Howardhuang98/rl_method

updated

README.md

@@ -8,6 +8,7 @@ School: Tianjin University, Priceless Lab
 ## Here you can use:
 * greedy hill climb
 * dynamic program: shortest path perspective
+* PC algorithm
 
 ## Project structure
 

datasets/Asian expert.csv

@@ -0,0 +1,9 @@
+,smoke,bronc,lung,asia,tub,either,dysp,xray
+smoke,0,0.9,0.9,0,0,0,0,0
+bronc,0,0,0,0,0,0,0.9,0
+lung,0,0,0,0,0,0.9,0,0
+asia,0,0,0,0,0.9,0,0,0
+tub,0,0,0,0,0,0.9,0,0
+either,0,0,0,0,0,0,0.9,0.9
+dysp,0,0,0,0,0,0,0,0
+xray,0,0,0,0,0,0,0,0

dlbn/base.py

@@ -0,0 +1,92 @@
+"""
+base class
+"""
+
+from abc import ABC
+from abc import abstractmethod
+
+import networkx as nx
+import numpy as np
+import pandas as pd
+from matplotlib import pyplot as plt
+
+
+
+class Estimator(ABC):
+
+    def load_data(self, data):
+        """
+        加载数据
+        """
+
+        if isinstance(data, pd.DataFrame):
+            self.data = data
+        elif isinstance(data, np.ndarray):
+            data = pd.DataFrame(data, columns=[range(data.shape[0])])
+            self.data = data
+        else:
+            raise ValueError("Data loading error")
+
+    def show_est(self):
+        print("=========Estimator Information=========")
+        print('''
+        ·▄▄▄▄    ▄▄▌    ▄▄▄▄·    ▐ ▄ 
+        ██▪ ██   ██•    ▐█ ▀█▪  •█▌▐█
+        ▐█· ▐█▌  ██▪    ▐█▀▀█▄  ▐█▐▐▌
+        ██. ██   ▐█▌▐▌  ██▄▪▐█  ██▐█▌
+        ▀▀▀▀▀•   .▀▀▀   ·▀▀▀▀   ▀▀ █▪
+        ''')
+        print(self.data.head(5))
+        print("Recover the BN with {} variables".format(len(self.data.columns)))
+
+    @abstractmethod
+    def run(self):
+        """
+        run the estimator
+        """
+
+    def show(self,):
+        if self.result_dag:
+            plt.figure()
+            nx.draw_networkx(self.result_dag)
+            plt.title("Bayesian network")
+            plt.show()
+        else:
+            raise ValueError("No result obtained")
+
+
+class Score(ABC):
+    """
+    Score base class
+    """
+
+    def __init__(self, data: pd.DataFrame):
+        self.data = data
+        self.state_names = {}
+        for var in list(data.columns.values):
+            self.state_names[var] = sorted(list(self.data.loc[:, var].unique()))
+        self.contingency_table = None
+
+    @abstractmethod
+    def local_score(self, x, parents):
+        """
+        return local score
+        """
+
+    def all_score(self, dag, detail=True):
+        """
+        return score on the DAG
+        """
+        score_dict = {}
+        score_list = []
+        for node in dag.nodes:
+            parents = list(dag.predecessors(node))
+            local_score = self.local_score(node, parents)
+            score_list.append(local_score)
+            if detail:
+                score_dict[node] = local_score
+        if detail:
+            return sum(score_list), score_dict
+        return sum(score_list)
+
+