fixed some issues in stock price prediction tutorial

x4nth055 · x4nth055 · commit 71fb55c720fa · 2020-09-03T16:43:16.000+02:00
diff --git a/machine-learning/stock-prediction/stock_prediction.ipynb b/machine-learning/stock-prediction/stock_prediction.ipynb
@@ -159,11 +159,10 @@
     "            sequence_data.append([np.array(sequences), target])\n",
     "\n",
     "    # get the last sequence by appending the last `n_step` sequence with `lookup_step` sequence\n",
-    "    # for instance, if n_steps=50 and lookup_step=10, last_sequence should be of 59 (that is 50+10-1) length\n",
-    "    # this last_sequence will be used to predict in future dates that are not available in the dataset\n",
+    "    # for instance, if n_steps=50 and lookup_step=10, last_sequence should be of 60 (that is 50+10) length\n",
+    "    # this last_sequence will be used to predict future stock prices not available in the dataset\n",
     "    last_sequence = list(sequences) + list(last_sequence)\n",
-    "    # shift the last sequence by -1\n",
-    "    last_sequence = np.array(pd.DataFrame(last_sequence).shift(-1).dropna())\n",
+    "    last_sequence = np.array(last_sequence)\n",
     "    # add to result\n",
     "    result['last_sequence'] = last_sequence\n",
     "    \n",
@@ -1053,9 +1052,9 @@
     "    return accuracy_score(y_test, y_pred)\n",
     "\n",
     "\n",
-    "def predict(model, data, classification=False):\n",
+    "def predict(model, data):\n",
     "    # retrieve the last sequence from data\n",
-    "    last_sequence = data[\"last_sequence\"][:N_STEPS]\n",
+    "    last_sequence = data[\"last_sequence\"][-N_STEPS:]\n",
     "    # retrieve the column scalers\n",
     "    column_scaler = data[\"column_scaler\"]\n",
     "    # reshape the last sequence\n",
diff --git a/machine-learning/stock-prediction/stock_prediction.py b/machine-learning/stock-prediction/stock_prediction.py
@@ -78,11 +78,10 @@ def load_data(ticker, n_steps=50, scale=True, shuffle=True, lookup_step=1,
             sequence_data.append([np.array(sequences), target])
 
     # get the last sequence by appending the last `n_step` sequence with `lookup_step` sequence
-    # for instance, if n_steps=50 and lookup_step=10, last_sequence should be of 59 (that is 50+10-1) length
-    # this last_sequence will be used to predict in future dates that are not available in the dataset
+    # for instance, if n_steps=50 and lookup_step=10, last_sequence should be of 60 (that is 50+10) length
+    # this last_sequence will be used to predict future stock prices not available in the dataset
     last_sequence = list(sequences) + list(last_sequence)
-    # shift the last sequence by -1
-    last_sequence = np.array(pd.DataFrame(last_sequence).shift(-1).dropna())
+    last_sequence = np.array(last_sequence)
     # add to result
     result['last_sequence'] = last_sequence
     
diff --git a/machine-learning/stock-prediction/test.py b/machine-learning/stock-prediction/test.py
@@ -28,9 +28,9 @@ def get_accuracy(model, data):
     return accuracy_score(y_test, y_pred)
 
 
-def predict(model, data, classification=False):
+def predict(model, data):
     # retrieve the last sequence from data
-    last_sequence = data["last_sequence"][:N_STEPS]
+    last_sequence = data["last_sequence"][-N_STEPS:]
     # retrieve the column scalers
     column_scaler = data["column_scaler"]
     # reshape the last sequence
