%PDF-1.7
%
1 0 obj
<>/Metadata 330 0 R/Pages 2 0 R/StructTreeRoot 69 0 R/Type/Catalog/ViewerPreferences 328 0 R>>
endobj
330 0 obj
<>stream
Time Series Forecasting Evaluation Analytics Data Science
Microsoft® Word 2016
When predicting across time, typical methodologies of prediction evaluation no longer hold true. It is not practical to take a holdout sample randomly from observations in the data set or even to use a typical k-fold cross-validation structure. Even newer methods of prediction evaluation in cross-sectional data like target shuffling should not just be applied to data where a temporal structure is inherent. How then can we determine if we have a good forecast or if we have reached our conclusions by random chance? This talk highlights advantages and disadvantages to techniques for evaluating predictions when forecasting future observations. It also discusses possible biases arising from time structures of data that should be considered. <br/><br/>Aric Labarr, Elder Research Inc.
Session 1862
en
jeff.foxx@sas.com
When predicting across time, typical methodologies of prediction evaluation no longer hold true. It is not practical to take a holdout sample randomly from observations in the data set or even to use a typical k-fold cross-validation structure. Even newer methods of prediction evaluation in cross-sectional data like target shuffling should not just be applied to data where a temporal structure is inherent. How then can we determine if we have a good forecast or if we have reached our conclusions by random chance? This talk highlights advantages and disadvantages to techniques for evaluating predictions when forecasting future observations. It also discusses possible biases arising from time structures of data that should be considered. <br/><br/>Aric Labarr, Elder Research Inc.
Microsoft® Word 2016
2018-03-11T23:24:39.000-04:00
2018-03-11T23:24:39.000-04:00
2018-04-03T15:45:22.331-04:00
1acdd0cccb4d633b1f203cf28cfde944a4458d09
327624
application/pdf
2018-04-03T15:45:22.409-04:00
Aric LaBarr
How to Evaluate Time Series Models
How Good is That Forecast? The Nuances of Prediction Evaluation Across Time
uuid:DCDC5D79-1D86-4FA8-8043-D2608A737AED
uuid:DCDC5D79-1D86-4FA8-8043-D2608A737AED
thirdparty
support:sgf-papers
year:2018
software:ETS
support:sgf-papers/session-type/breakout
support:sgf-papers/topic/analytics/forecasting-econometrics
support:customer-roles/data-scientist
endstream
endobj
2 0 obj
<>
endobj
69 0 obj
<>
endobj
328 0 obj
<>
endobj
71 0 obj
<>
endobj
70 0 obj
<>
endobj
75 0 obj
[74 0 R 77 0 R 78 0 R 79 0 R 80 0 R 81 0 R 82 0 R 83 0 R 84 0 R 85 0 R 86 0 R 87 0 R 88 0 R 89 0 R]
endobj
91 0 obj
[90 0 R 92 0 R 93 0 R 94 0 R 95 0 R 96 0 R 98 0 R 101 0 R 102 0 R 103 0 R 104 0 R 106 0 R 107 0 R 105 0 R]
endobj
99 0 obj
<>
endobj
109 0 obj
[108 0 R 110 0 R 112 0 R 115 0 R 116 0 R 118 0 R 121 0 R 122 0 R 128 0 R 131 0 R 132 0 R 133 0 R 134 0 R 187 0 R 188 0 R 190 0 R 191 0 R 139 0 R 139 0 R 141 0 R 141 0 R 143 0 R 143 0 R 147 0 R 147 0 R 149 0 R 149 0 R 151 0 R 151 0 R 154 0 R 156 0 R 158 0 R 161 0 R 161 0 R 163 0 R 163 0 R 165 0 R 165 0 R 168 0 R 170 0 R 172 0 R 175 0 R 175 0 R 177 0 R 177 0 R 179 0 R 179 0 R 182 0 R 184 0 R 186 0 R 126 0 R 125 0 R 124 0 R]
endobj
113 0 obj
<>
endobj
119 0 obj
<>
endobj
129 0 obj
<>
endobj
193 0 obj
[192 0 R 195 0 R 198 0 R 199 0 R 204 0 R 205 0 R 206 0 R 207 0 R 208 0 R 212 0 R 200 0 R 203 0 R 202 0 R]
endobj
196 0 obj
<>
endobj
217 0 obj
[216 0 R 219 0 R 221 0 R 223 0 R 226 0 R 231 0 R 232 0 R 233 0 R 237 0 R 238 0 R 240 0 R 243 0 R 244 0 R 249 0 R 248 0 R 227 0 R 236 0 R 235 0 R 247 0 R 246 0 R]
endobj
229 0 obj
<>
endobj
241 0 obj
<>
endobj
251 0 obj
[250 0 R 253 0 R 256 0 R 257 0 R 259 0 R 260 0 R 264 0 R 265 0 R 267 0 R 270 0 R 271 0 R 276 0 R 277 0 R 278 0 R 258 0 R 263 0 R 262 0 R 275 0 R 274 0 R 273 0 R]
endobj
254 0 obj
<>
endobj
268 0 obj
<>
endobj
281 0 obj
[280 0 R 284 0 R 285 0 R 290 0 R 291 0 R 292 0 R 293 0 R 294 0 R 298 0 R 300 0 R 301 0 R 302 0 R 303 0 R 304 0 R 305 0 R 309 0 R 310 0 R 311 0 R 312 0 R 289 0 R 288 0 R 287 0 R]
endobj
282 0 obj
<>
endobj
307 0 obj
<>
endobj
306 0 obj
<>
endobj
53 0 obj
<>/MediaBox[0 0 612 792]/Parent 2 0 R/Resources<>/Font<>/ProcSet[/PDF/Text/ImageB/ImageC/ImageI]/XObject<>>>/StructParents 15/Tabs/S/Type/Page>>
endobj
54 0 obj
<>stream
x\Ys8~wMڊ`7S)GfknmM%1jRxDҲldjbht}@'K>ūW':-䬪'?VfY*O>o5^zSUu>=g'=/SD&/*㣳㣓7Jdg