modify

stock/dnn_predict_dmi.py

@@ -17,8 +17,28 @@ def read_data(path):
     train_y=[s[size-1] for s in lines]
     return np.array(train_x),np.array(train_y),lines
 
+def _score(fact, line):
+    with open('dnn_predict_dmi_18d.txt', 'a') as f:
+        f.write(str([line[-2], line[-1]]) + "\n")
 
-def predict(file_path='', model_path='15min_dnn_seq.h5'):
+    up_right = 0
+    up_error = 0
+
+    if fact[0] == 1:
+        up_right = up_right + 1.12
+    elif fact[1] == 1:
+        up_right = up_right + 1.06
+    elif fact[2] == 1:
+        up_right = up_right + 1
+    elif fact[3] == 1:
+        up_right = up_right + 0.94
+    else:
+        up_error = up_error + 1
+        up_right = up_right + 0.88
+    return up_right,up_error
+
+
+def predict(file_path='', model_path='15min_dnn_seq.h5', idx=-1):
     test_x,test_y,lines=read_data(file_path)
 
     model=load_model(model_path)
@@ -34,58 +54,31 @@ def predict(file_path='', model_path='15min_dnn_seq.h5'):
     i = 0
     result=model.predict(test_x)
     win_dnn = []
-    with open('dnn_predict_dmi_18d.txt', 'a') as f:
-        for r in result:
-            fact = test_y[i]
-            if r[0] > 0.5:
-                f.write(str([lines[i][-2], lines[i][-1]]) + "\n")
-                win_dnn.append([lines[i][-2], lines[i][-1]])
-                if fact[0] == 1:
-                    up_right = up_right + 1.12
-                elif fact[1] == 1:
-                    up_right = up_right + 1.06
-                elif fact[2] == 1:
-                    up_right = up_right + 1
-                elif fact[3] == 1:
-                    up_right = up_right + 0.94
-                else:
-                    up_error = up_error + 1
-                    up_right = up_right + 0.88
+    for r in result:
+        fact = test_y[i]
+
+        if idx in [0]:
+            if r[0] > 0.5 or r[1] > 0.5:
+                pass
+                # if fact[0] == 1:
+                #     up_right = up_right + 1.12
+                # elif fact[1] == 1:
+                #     up_right = up_right + 1.06
+                # elif fact[2] == 1:
+                #     up_right = up_right + 1
+                # elif fact[3] == 1:
+                #     up_right = up_right + 0.94
+                # else:
+                #     up_error = up_error + 1
+                #     up_right = up_right + 0.88
+                # up_num = up_num + 1
+        else:
+            if r[0] > 0.5 or r[1] > 0.5:
+                tmp_right,tmp_error = _score(fact, lines[i])
+                up_right = tmp_right + up_right
+                up_error = tmp_error + up_error
                 up_num = up_num + 1
-            elif r[1] > 0.5:
-                f.write(str([lines[i][-2], lines[i][-1]]) + "\n")
-                win_dnn.append([lines[i][-2], lines[i][-1]])
-                if fact[0] == 1:
-                    up_right = up_right + 1.12
-                elif fact[1] == 1:
-                    up_right = up_right + 1.06
-                elif fact[2] == 1:
-                    up_right = up_right + 1
-                elif fact[3] == 1:
-                    up_right = up_right + 0.94
-                else:
-                    up_error = up_error + 1
-                    up_right = up_right + 0.88
-                up_num = up_num + 1
-
-            if r[3] > 0.6:
-                f.write(str([lines[i][-2], lines[i][-1]]) + "\n")
-                win_dnn.append([lines[i][-2], lines[i][-1]])
-                if fact[0] == 1:
-                    down_error = down_error + 1
-                    down_right = down_right + 1.12
-                elif fact[1] == 1:
-                    down_right = down_right + 1.06
-                elif fact[2] == 1:
-                    down_right = down_right + 1
-                elif fact[3] == 1:
-                    down_right = down_right + 0.94
-                else:
-                    down_right = down_right + 0.88
-                down_num = down_num + 1
-            elif r[4] > 0.6:
-                f.write(str([lines[i][-2], lines[i][-1]]) + "\n")
-                win_dnn.append([lines[i][-2], lines[i][-1]])
+            elif r[3] > 0.5 or r[4] > 0.5:
                 if fact[0] == 1:
                     down_error = down_error + 1
                     down_right = down_right + 1.12
@@ -99,9 +92,11 @@ def predict(file_path='', model_path='15min_dnn_seq.h5'):
                     down_right = down_right + 0.88
                 down_num = down_num + 1
 
-            i = i + 1
+        i = i + 1
     if up_num == 0:
         up_num = 1
+    if down_num == 0:
+        down_num = 1
     print('DNN', up_right, up_num, up_right/up_num, up_error/up_num, down_right/down_num, down_error/down_num)
     return win_dnn,up_right/up_num,down_right/down_num
 
@@ -110,11 +105,13 @@ def multi_predict():
     r = 0;
     p = 0
     # for x in range(0, 12): # 0,2,3,4,6,8,9,10,11
-    # for x in [5,6,11]:
-    for x in [2,4,7,10]: # 2表现最好 优秀的
+    # for x in [2,3,4,5,6,7,8,9,11]: 10_18,0没数据需要重新计算
+    for x in [0,1,10]:
+    # for x in [2,4,7,10]: # 2表现最好 优秀的 0,8正确的反向指标,(9错误的反向指标 样本量太少)
         print(x)
     # for x in [0,2,5,6,7]: # 5表现最好
-        win_dnn, up_ratio,down_ratio = predict(file_path='D:\\data\\quantization\\kmeans\\stock9_18_test_' + str(x) + '.log', model_path='18d_dnn_seq_' + str(x) + '.h5')
+        win_dnn, up_ratio,down_ratio = predict(file_path='D:\\data\\quantization\\kmeans\\stock10_18_test_' + str(x) + '.log',
+                                               model_path='18d_dnn_seq_' + str(x) + '.h5', idx=x)
         r = r + up_ratio
         p = p + down_ratio
     print(r, p)
@@ -132,9 +129,9 @@ industry = ['全国地产', '区域地产', '酒店餐饮',
             '塑料', '电器连锁', '半导体', '乳制品',]
 
 
-def predict_today(day):
+def predict_today(day, model='10_18d'):
     lines = []
-    with open('D:\\data\\quantization\\stock9_18_' +  str(day) +'.log') as f:
+    with open('D:\\data\\quantization\\stock' + model + '_' +  str(day) +'.log') as f:
         for line in f.readlines()[:]:
             line = eval(line.strip())
             if line[-1][0].startswith('0') or line[-1][0].startswith('3'):
@@ -148,9 +145,9 @@ def predict_today(day):
 
     models = []
     for x in range(0, 12):
-        models.append(load_model('18d_dnn_seq_' + str(x) + '.h5'))
+        models.append(load_model(model + '_dnn_seq_' + str(x) + '.h5'))
 
-    x = 21 # 每条数据项数
+    x = 24 # 每条数据项数
     k = 18 # 周期
     for line in lines:
         v = line[1:x*k + 1]
@@ -161,21 +158,21 @@ def predict_today(day):
         # print(v)
         r = estimator.predict(v)
 
-        if r[0] in [5,6,11]:
-            train_x = np.array([line[:size - 1]])
-
-            result = models[r[0]].predict(train_x)
-            if result[0][3] > 0.5 or result[0][4] > 0.5:
-                stock = code_table.find_one({'ts_code':line[-1][0]})
-                if stock['name'].startswith('ST') or stock['name'].startswith('N') or stock['name'].startswith('*'):
-                    continue
-                if line[0] > 80:
-                    continue
-                if stock['industry'] in industry:
-                    pass
-                    # print(line[-1], stock['name'], stock['industry'], 'sell')
-
-        if r[0] in [2,4,7,10]:
+        # if r[0] in [1,6,10]:
+        #     train_x = np.array([line[:size - 1]])
+        #
+        #     result = models[r[0]].predict(train_x)
+        #     if result[0][3] > 0.5 or result[0][4] > 0.5:
+        #         stock = code_table.find_one({'ts_code':line[-1][0]})
+        #         if stock['name'].startswith('ST') or stock['name'].startswith('N') or stock['name'].startswith('*'):
+        #             continue
+        #         if line[0] > 80:
+        #             continue
+        #         if stock['industry'] in industry:
+        #             pass
+        #             # print(line[-1], stock['name'], stock['industry'], 'sell')
+
+        if r[0] in [2,3,4,5,6,7,8,9,11]:
             train_x = np.array([line[:size - 1]])
 
             result = models[r[0]].predict(train_x)
@@ -198,12 +195,12 @@ def predict_today(day):
                     continue
 
                 # 指定某几个行业
-                # if stock['industry'] in industry:
-                print(line[-1], stock['name'], stock['industry'], 'buy')
+                if stock['industry'] in industry:
+                    print(line[-1], stock['name'], stock['industry'], 'buy')
 
 
 if __name__ == '__main__':
     # predict(file_path='D:\\data\\quantization\\stock6_5_test.log', model_path='5d_dnn_seq.h5')
     # predict(file_path='D:\\data\\quantization\\stock6_test.log', model_path='15m_dnn_seq.h5')
-    multi_predict()
-    # predict_today(20200219)
+    # multi_predict()
+    predict_today(20200221)

stock/dnn_train_dmi.py

@@ -36,12 +36,14 @@ def read_data(path):
 
     return X_resampled,y_resampled,np.array(test_x),np.array(test_y)
 
+data_dir = 'D:\\data\\quantization\\kmeans\\'
 
-def resample(path):
+
+def resample(path, suffix='test'):
     lines = []
     with open(path) as f:
         i = 0
-        for x in range(110000):
+        for x in range(67000): # 42万 10万 6.7万
             # print(i)
             lines.append(eval(f.readline().strip()))
             i = i + 1
@@ -49,9 +51,9 @@ def resample(path):
 
     file_list = []
     for x in range(0, 12):
-        file_list.append(open('D:\\data\\quantization\\kmeans\\stock9_18_train_' + str(x) + '.log', 'a'))
+        file_list.append(open(data_dir + 'stock11_18d_' + suffix + '_' + str(x) + '.log', 'a'))
 
-    x = 21 # 每条数据项数
+    x = 24 # 每条数据项数
     k = 18 # 周期
     for line in lines:
         v = line[1:x*k + 1]
@@ -64,14 +66,14 @@ def resample(path):
         file_list[r[0]].write(str(line) + '\n')
 
 
-def mul_train():
-    # for x in range(0, 12):
-    for x in [11,0,1,3,8,9]:
-    # for x in [2,4,7,10]:
-        score = train(input_dim=384, result_class=5, file_path="D:\\data\\quantization\\kmeans\\stock9_18_train_" + str(x) + ".log",
+def mul_train(name="10_18"):
+    for x in range(0, 12):
+    # for x in [11,0,1,3,8,9]:
+    # for x in [11,0,1,3,5,6,8,9]:
+        score = train(input_dim=440, result_class=5, file_path=data_dir + "stock"+ name + "_train_" + str(x) + ".log",
               model_name='18d_dnn_seq_' + str(x) + '.h5')
 
-        with open('D:\\data\\quantization\\kmeans\\stock9_18_dmi.log', 'a') as f:
+        with open(data_dir + 'stock' + name + '_dmi.log', 'a') as f:
             f.write(str(x) + ':' + str(score[1]) + '\n')
 
 
@@ -89,7 +91,7 @@ def train(input_dim=400, result_class=3, file_path="D:\\data\\quantization\\stoc
     model.add(Dense(units=120 + input_dim, activation='relu'))
     model.add(Dropout(0.2))
     model.add(Dense(units=120+input_dim, activation='selu'))
-    model.add(Dropout(0.2))
+    model.add(Dropout(0.1))
     model.add(Dense(units=120+input_dim, activation='selu'))
     model.add(Dense(units=512, activation='relu'))
 
@@ -97,7 +99,7 @@ def train(input_dim=400, result_class=3, file_path="D:\\data\\quantization\\stoc
     model.compile(loss='categorical_crossentropy', optimizer="adam",metrics=['accuracy'])
 
     print("Starting training ")
-    model.fit(train_x, train_y, batch_size=4096, epochs=900 + 6*int(len(train_x)/600), shuffle=True)
+    model.fit(train_x, train_y, batch_size=4096, epochs=555 + 5*int(len(train_x)/888), shuffle=True)
     score = model.evaluate(test_x, test_y)
     print(score)
     print('Test score:', score[0])
@@ -115,5 +117,5 @@ def train(input_dim=400, result_class=3, file_path="D:\\data\\quantization\\stoc
 
 
 if __name__ == '__main__':
-    # resample('D:\\data\\quantization\\stock9_18_1.log')
-    mul_train()
+    # resample('D:\\data\\quantization\\stock11_18d_test.log', suffix='test')
+    mul_train('11_18d')