Forecasting Software Survey

For this year's forecasting software survey, as in the past, OR/MS Today attempted to include as many forecasting products as possible. We contacted recent survey participants, as well as any new vendors that came to our attention through the authors or our own contacts. We asked each respondent to complete an online questionnaire covering a comprehensive list of questions spanning features and capabilities, recent enhancements, licensing and fees, technical support and other areas. OR/MS Today followed up multiple times with each vendor to help ensure that we obtained as many responses as possible. Vendors not included in this issue, of which there are many, are invited to submit a completed online questionnaire, and their product will be added to the online version of the forecasting survey. The survey questionnaire is available at:
https://www.informs.org/ORMS-Today/2018-OR-MS-Today-Forecasting-Software-Survey

The purpose of the survey is simply to inform the reader of what is available. The information in the survey comes directly from the vendors, and no attempt was made to verify or validate the information they gave us.

Be sure to read the accompanying article.

The survey is divided into 28 separate pages. Following is an index of the pages and the information they contain:

Page 1
Vendor
Supported operating system(s): Windows, macOS, Linux, Others
Computer architecture(s) supported: 32-bit, 64-bit, Browser / cloud-based

Page 2
Data Handling:
- Import: Reads CSV files; Supports database connectivity; Supports cloud connectivity; Is there a maximum number of observations in time-series?; If yes, how many?
- Reads Excel files automatically (besides "cut-paste"): Reads Excel files automatically (besides "cut-paste")?; No, does not reads Excel files automatically; If not, what modifications must be made? (i.e., user must save as earlier version, insert additional information in certain rows and/or columns, etc.)

Page 3
Data Handling (continued):
- Export: Exports output to Excel file; Exports output to CSV file; Exports session output (i.e., entire output can be saved as a Word, RTF file, etc.); Specify formats

Page 4
Data Handling (continued):
- Supports third-party libraries to extend capability
- Stand-alone
- Integrates with other programming languages
- Specify other programming languages
- Which third-party software does software integrate with?

Page 5
Forecasting Features:
- General Features:
Not Applicable, Allows selection of estimation and hold-out sample, Fixed horizon evaluations, Rolling origin evaluations, Multiple error measures, Benchmark forecasts (e.g. Random Walk)

Page 6
Forecasting Features (continued):
- General Features (continued): Saving forecasts and forecast errors for further analysis, Model combination, Batch forecasting for multiple series, Supports multiple seasonalities, Prediction intervals after saving forecasts

Page 7
Forecasting Features (continued):
- Hierarchical Forecasting: None, Top down, Bottom-Up, Multiple hierarchies, Optimal combination reconciliation

Page 8
Forecasting Features (continued):
- Temporal Aggregation: None, Different data frequencies (e.g. daily, monthly), Multiple (e.g. Temporal Hierarchies, MAPA)
- Judgmental Forecasting: None, Judgmental adjustment of statistical forecasts, Adjustment stored in forecast history

Page 9
Exploratory Analysis and Graphics:
- Graphical and Exploratory Analysis: None, Time-series Plot, Seasonal Plot, Scatter Plot, Multiple Scatterplot

Page 10
Exploratory Analysis and Graphics (continued):
- Graphical and Exploratory Analysis (continued): Autocorrelation Function; Autocorrelation, Partial; Bar, Pie, Other Business Chart; Box Plots; Time Series Transformations; Simple Moving Averages and Smoothing

Page 11
Exploratory Analysis and Graphics (continued):
- Decomposition: None; Classical Decomposition; Automatic outlier adjustment; Census Bureau Methods (i.e., X-12, X-13-ARIMA); Others

Page 12
Forecasting Methods Available
- Exponential Smoothing: Not included, Simple Moving Average, Simple (one parameter) Exponential Smoothing, Holt's Two-Parameter (including trend), Winters' Three-Parameter Methods (including seasonal), Damped Trend

Page 13
Forecasting Methods Available (continued):
- Exponential Smoothing (continued): Adaptive Response Rate; TBATS (including multiple seasonalities using trigonometric functions, and Autoregressive errors); Exponential Smoothing in State Space Form and Estimation; Automatic Model Selection; Do all methods offer optimal estimation of parameter; If methods do not offer optimal estimation of parameters please comment

Page 14
Forecasting Methods Available (continued):
- Methods for Intermittent Demand: Not Included; Croston; Syntetos Boylan Approximation; Teunter-Syntetos Babai; Other Methods for Intermittent Demand; Specify other Methods for Intermittent Demand; Other features, please specify

Page 15
Forecasting Methods Available (continued):
- ARIMA (Box-Jenkins and ARCH/GARCH): Not Included; Regular ARIMA; Seasonal ARIMA; ARCH/ GARCH heteroscedastic modelling

Page 16
Forecasting Methods Available (continued):
- Regression Models: Not Included; Multiple linear regression; Prediction intervals; Multicollinearity statistics; Outlier/ Leverage identification; General autocorrelation tests of residuals

Page 17
Forecasting Methods Available (continued):
- Regression models (continued): Extended range of tests for misspecification (e.g. heteroscedasticity); Stability of coefficient tests; Maximum number of independent variables; How many Maximum number of independent variables; Is there a maximum number of observations?; How many is the maximum number of observations?

Page 18
Forecasting Methods Available (continued):
- Extended Range of Regression Models: Not Included; Dynamic regression (e.g. autoregressive distributed lag models with lags in both dependent and explanatory variables); Non-linear regression models; Lasso (or Ridge) regression; Logistic Regression

Page 19
Forecasting Methods Available (continued):
Extended range of regression models (continued): Time-varying parameter regression; Model selection (e.g. forward selection AIC, best subsets, etc.); Other regression models; Specify other regression models

Page 20
Forecasting Methods Available (continued):
- Growth and Diffusion Models: Not Included; Exponential; Bass; Weibull; Logistic; Gompertz; Other Growth and Diffusion Models; Specify other Growth and Diffusion Models

Page 21
Forecasting Methods Available (continued):
Machine Learning - Neural Networks: Not Included; Feed-forward; Recurrent; Deep-learning; Use ensembles; Support automatic specification

Page 22
Forecasting Methods Available (continued):
Machine Learning: Not Included; Support Vector Machine; K-Nearest Neighbors; Gradient Boosting; Random Forrest; Classification and Regression Trees; Other machine-learning techniques; Specify other machine-learning techniques

Page 23
Forecasting Methods Available (continued):
- Additional Capabilities: None; Spectral / Fourier Analysis; State Space Models; Other Additional Capabilities; Specify Additional Capabilities

Page 24
Documentation:
- Support Available: None; Online help and tutorials; Online Chat / Instant Messenger; Online community support (e.g. Forum, StackOverflow); Other features (e.g., software offers forecasting advice, software statistically validates the model etc.); Specify other support features

Page 25
Pricing:
License Type and Evaluation: None; Commercial license; Educational license; Can user download a "trial version" of the software; Pricing model (i.e. one-time fee, subscription based); Specify pricing model

Page 26
New features and enhancements since June 2016
Comments

Page 27
Describe recent trends or market demands that are reshaping the forecasting software space
How do you view the increased adoption of open-source forecasting tools? How is it affecting your company and the industry?

Vendor Directory