Added case 4: w/ static from both regions

nipunbatra · Jan 5, 2017 · f6054c8 · f6054c8
1 parent 11dffe9
commit f6054c8
Show file tree

Hide file tree

Showing 4 changed files with 38 additions and 63 deletions.
diff --git a/mf_approach/code/both_regions_dd.py b/mf_approach/code/both_regions_dd.py
@@ -1,7 +1,8 @@
 ###################CASES#######################################
 ### CASE 1: Train and test on homes from SD w/o static features
 ### CASE 2: Train on homes from SD and Austin w/o static features
-### Case 3: Train on homes from SD and Austin w/ temp feature
+### Case 3: Train on homes from SD and Austin w/ static and temp feature
+### Case 4: Train and test on homes from SD w static and temp features
 ################################################################
 
 
@@ -76,6 +77,14 @@ def create_region_df(region):
         aus_dfc[key]=dds['Austin'][key_num]
     df = pd.concat([sd_df, aus_df])
     dfc = pd.concat([sd_dfc, aus_dfc])
+elif case==4:
+    df = sd_df
+    dfc = sd_dfc
+    for key_num, key in enumerate(dd_keys):
+        sd_df[key]=dds['SanDiego'][key_num]
+        sd_dfc[key]=dds['SanDiego'][key_num]
+
+
 
 
 
@@ -107,7 +116,7 @@ def create_region_df(region):
     start, end=1,13
 X_matrix, X_normalised, col_max, col_min, appliance_cols, aggregate_cols = preprocess(df, dfc, appliance)
 
-if case==3:
+if case>=3:
     static_features = get_static_features_region_level(dfc, X_normalised)
 else:
     static_features = get_static_features(dfc, X_normalised)
@@ -119,7 +128,7 @@ def create_region_df(region):
 #all_feature_homes = dfc[(dfc.full_agg_available == 1) & (dfc.md_available == 1)][all_cols].dropna().index
 
 
-if case==3:
+if case>=3:
     max_f = 20
 else:
     max_f=1

diff --git a/mf_approach/notebooks/final-all-homes.ipynb b/mf_approach/notebooks/final-all-homes.ipynb
@@ -957,8 +957,9 @@
   }
  ],
  "metadata": {
+  "anaconda-cloud": {},
   "kernelspec": {
-   "display_name": "Python 2",
+   "display_name": "Python [default]",
    "language": "python",
    "name": "python2"
   },
@@ -972,7 +973,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.11"
+   "version": "2.7.12"
   }
  },
  "nbformat": 4,

diff --git a/mf_approach/notebooks/kdd17.ipynb b/mf_approach/notebooks/kdd17.ipynb
@@ -4345,9 +4345,7 @@
      "output_type": "execute_result"
     }
    ],
-   "source": [
-    ""
-   ]
+   "source": []
   },
   {
    "cell_type": "code",
@@ -4356,31 +4354,29 @@
     "collapsed": true
    },
    "outputs": [],
-   "source": [
-    ""
-   ]
+   "source": []
   }
  ],
  "metadata": {
   "anaconda-cloud": {},
   "kernelspec": {
-   "display_name": "Python 2",
+   "display_name": "Python [default]",
    "language": "python",
    "name": "python2"
   },
   "language_info": {
    "codemirror_mode": {
     "name": "ipython",
-    "version": 2.0
+    "version": 2
    },
    "file_extension": ".py",
    "mimetype": "text/x-python",
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.11"
+   "version": "2.7.12"
   }
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
diff --git a/mf_approach/notebooks/main-result-sd.ipynb b/mf_approach/notebooks/main-result-sd.ipynb
@@ -49,7 +49,7 @@
    },
    "outputs": [],
    "source": [
-    "mf_pred = pickle.load(open('../data/mf_pred_all_homes.pkl','r'))\n",
+    "mf_pred = pickle.load(open('../data/out_3_dd.pkl','r'))\n",
     "knn_results = pickle.load(open('../data/knn_pred_all_homes.pkl','r'))"
    ]
   },
@@ -66,14 +66,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": 8,
    "metadata": {
     "collapsed": true
    },
    "outputs": [],
    "source": [
     "def compute_rmse_fraction(appliance, pred_df):\n",
-    "    region = \"Austin\"\n",
+    "    region = \"SanDiego\"\n",
     "    df = out_overall[region]\n",
     "    #pred_df = mf_pred[appliance][appliance_feature][latent_factors]\n",
     "    gt_df = df[pred_df.columns].ix[pred_df.index]\n",
@@ -89,8 +89,8 @@
     "    \n",
     "    pred_fraction = pred_fraction.dropna()\n",
     "    gt_fraction = gt_fraction.dropna()\n",
-    "    print pred_fraction.describe()\n",
-    "    print gt_fraction.describe()\n",
+    "    #print pred_fraction.describe()\n",
+    "    #print gt_fraction.describe()\n",
     "\n",
     "\n",
     "    rms = np.sqrt(mean_squared_error(pred_fraction,gt_fraction))\n",
@@ -99,44 +99,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 10,
    "metadata": {
     "collapsed": true
    },
    "outputs": [],
    "source": [
-    "result = {'MF':{},'KNN':{},'FHMM':{}}"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Computing for KNN"
+    "result = {'MF':{}}"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
-   "metadata": {
-    "collapsed": false
-   },
-   "outputs": [],
-   "source": [
-    "for appliance, appliance_df in knn_results.iteritems():\n",
-    "    \n",
-    "    pred_df = appliance_df\n",
-    "    pred_df.columns = [appliance+\"_\"+str(x) for x in pred_df.columns]\n",
-    "    rmse_fraction = compute_rmse_fraction(appliance, pred_df)\n",
-    "    \n",
-    "\n",
-    "    #rms = np.sqrt(mean_squared_error(gt_df, pred_df))\n",
-    "    result['KNN'][appliance]=rmse_fraction"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 11,
    "metadata": {
     "collapsed": false
    },
@@ -145,13 +119,9 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "hvac\n",
-      "wm\n",
       "fridge\n",
-      "mw\n",
-      "oven\n",
-      "dw\n",
-      "{'hvac': ('area', 9, 23.062443182449364), 'wm': ('occ', 3, 1.292907746177548), 'fridge': ('occ_area', 7, 56.419845955447094), 'mw': ('occ_area', 6, 0.9841175302519054), 'oven': ('occ', 8, 1.4241289192051436), 'dw': ('None', 5, 16.697917236289435)}\n"
+      "hvac\n",
+      "{'fridge': ('occ', 9, 13.431200129690193), 'hvac': ('dd_11', 3, 14.022029756205505)}\n"
      ]
     }
    ],
@@ -163,7 +133,7 @@
     "    for appliance_feature, appliance_feature_struct in appliance_struct.iteritems():\n",
     "        for latent_factors, appliance_df in appliance_struct[appliance_feature].iteritems():\n",
     "            try:\n",
-    "                region = \"Austin\"\n",
+    "                region = \"SanDiego\"\n",
     "                df = out_overall[region]\n",
     "                pred_df = mf_pred[appliance][appliance_feature][latent_factors]\n",
     "                \n",
@@ -732,30 +702,29 @@
     "collapsed": true
    },
    "outputs": [],
-   "source": [
-    ""
-   ]
+   "source": []
   }
  ],
  "metadata": {
+  "anaconda-cloud": {},
   "kernelspec": {
-   "display_name": "Python 2",
+   "display_name": "Python [default]",
    "language": "python",
    "name": "python2"
   },
   "language_info": {
    "codemirror_mode": {
     "name": "ipython",
-    "version": 2.0
+    "version": 2
    },
    "file_extension": ".py",
    "mimetype": "text/x-python",
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.11"
+   "version": "2.7.12"
   }
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}