@@ -24,7 +24,7 @@ Locally
24
24
25
25
pip install " oracle_ads[forecast]"
26
26
27
- ---
27
+ .. rst-class :: page-break
28
28
29
29
🚀 Getting Started
30
30
==================
@@ -53,7 +53,7 @@ Using the CLI
53
53
ads operator run -f forecast.yaml
54
54
55
55
56
- ---
56
+ .. rst-class :: page-break
57
57
58
58
Using the API
59
59
---------------
@@ -75,7 +75,7 @@ Using the API
75
75
result = operate(config)
76
76
77
77
78
- ---
78
+ .. rst-class :: page-break
79
79
80
80
Using the Notebook UI
81
81
------------------------
@@ -91,7 +91,7 @@ Simply fill in the fields and click "run":
91
91
92
92
.. image :: ./images/notebook_form_filled.png
93
93
94
- ---
94
+ .. rst-class :: page-break
95
95
96
96
🧠 Tweak the Model
97
97
===================
@@ -105,12 +105,12 @@ Select a specific model
105
105
name : arima
106
106
107
107
The model name can be any of the following:
108
- - **Prophet ** - Recommended for smaller datasets, and datasets with seasonality or holidays
109
- - **ARIMA ** - Recommended for highly cyclical datasets
110
- - **AutoMLx ** - Oracle Lab's proprietary modelling framework
111
- - **NeuralProphet ** - Recommended for large or wide datasets
112
- - **AutoTS ** - M6 Benchmark winner. Recommended if the other frameworks aren't providing enough accuracy
113
- - **Auto-Select ** - The best of all of the above. Recommended for comparing the above frameworks. Caution, it can be very slow.
108
+ - **Prophet ** - Recommended for smaller datasets, and datasets with seasonality or holidays
109
+ - **ARIMA ** - Recommended for highly cyclical datasets
110
+ - **AutoMLx ** - Oracle Lab's proprietary modelling framework
111
+ - **NeuralProphet ** - Recommended for large or wide datasets
112
+ - **AutoTS ** - M6 Benchmark winner. Recommended if the other frameworks aren't providing enough accuracy
113
+ - **Auto-Select ** - The best of all of the above. Recommended for comparing the above frameworks. Caution, it can be very slow.
114
114
115
115
116
116
Auto-Select the Best Model
@@ -163,7 +163,8 @@ With ``prophet``, for instance, there are options to dictate seasonality and cha
163
163
seasonality_mode : multiplicative
164
164
changepoint_prior_scale : 0.05
165
165
166
- ---
166
+
167
+ .. rst-class :: page-break
167
168
168
169
169
170
➕ Add Additional Column(s)
@@ -275,7 +276,7 @@ The store owner may also wish to run a multi-variate forecast and thus include
275
276
276
277
Notice that this additional_data would only be capable of forecasting a horizon of 1 (on 01-03-2024).
277
278
278
- ----
279
+ .. rst-class :: page-break
279
280
280
281
Sourcing Data
281
282
=================
@@ -393,7 +394,7 @@ When additional data is provided, the Operator can optionally generate explanati
393
394
target_column : y
394
395
generate_explanations : True
395
396
396
- ---
397
+ .. rst-class :: page-break
397
398
398
399
🧾 Disable File Generation
399
400
============================
@@ -405,7 +406,7 @@ When additional data is provided, the Operator can optionally generate explanati
405
406
generate_explanations_file : False
406
407
generate_metrics_file : False
407
408
408
- ---
409
+ .. rst-class :: page-break
409
410
410
411
📏 Change Evaluation Metric
411
412
============================
@@ -434,7 +435,7 @@ The metric can be optionally specified in the YAML file (default: "smape"):
434
435
target_column : y
435
436
metric : rmse
436
437
437
- ---
438
+ .. rst-class :: page-break
438
439
439
440
🧵 Run as a Job
440
441
============================
0 commit comments