Explore time series data like never before. Visualise all data in plots to understand the data and identify patterns of seasonality, trend, as well as intermittency and outliers. Use rigorous statistical tests to build up confidence in your analysis. Interactively transform the data, de-trend or de-seasonalise or both, remove outliers and missing values, and re-analyse your graphs and tests. iqast® will allow you to find insights fast.
Run a Pareto-Analysis across 1,000s of time series to determine how to best forecast each one. Best-practices suggest segmenting by relevance (ABC) and forecastability (XYZ) with flexible customisation or groups and metrics. Classify time series further by their time patterns (LTSI) to identify seasonal, trended, and intermittent series, and by lifecycle (NE) of new and end-of-life products. Interactively slice-and-dice all data across segments in multiple graphs, drilling down into individual time series, or exploring slow-moving items versus fast-movers for their forecast accuracy.
Cluster Analysis allows you to find similar time series in a large assortment of time series, or of similar calendar motifs, e.g. days of the week, in a single time series. Identify similar seasonal behavior across many series, or unusual versus usual Mondays within a single series. Interactively build clusters using various algorithms, and iteratively refine clusters by transforming and drilling deeper into you time series. Visualise all results using insightful graphs of cluster membership and purity.
Exploit the latest research in algorithms from artificial & computational intelligence, machine learning, data science, and predictive analytics. Utilize the nonlinear capabilities of Artificial Neural Networks, Support Vector Regression, k-Nearest Neighbours, and Decision Trees for forecasting. Compare these with benchmark algorithms from statistics, including Exponential Smoothing for single and multiple Seasonality, ARIMA methods, and various forms of Regression, to combine the best of all scientific disciplines in predictive modelling.
Utilize the latest research in forecasting model & profile selection. Employ rigorous statistical tests to pre-filter results, select using one of the 52 standard forecast error metrics or your custom-made one, using 1-step-ahead or trace-forecast errors for one or many customized horizons, for in-sample or out-of-sample evaluation.
iqast allows you to replicate all forecast models available in the software SAP® APO DP 3.1a or SAP® SCM 4.0, and therefore to externally simulate profile selection, suggest optimal profile and improve forecasting for SAP® APO DP.
Simply import using multiple standard import formats, including xls and csv files, SAP® master data etc. Using custom import and export wizards you can quickly define customised import formats and load your data, including multiple customized product hierarchies.
You can easily export actual forecasts for future and past time origins, selected models / profiles, ABC-XYZ indicators and all other features. Within the tool, export all grids, tables and graphs using Microsoft Windows® standard functionality. We even develop custom import and export filters for fully automatic data handling and integration into your computer systems.
Analyse and report forecasting accuracy across various interactive grids, tables and using graphs and dashboards. Select different errors, horizons, or time segments, sort and filter by ABC-XYZ or product hierarchies, or any of your imported product hierarchies. Drill into segments of interest, analyse which models were chosen and identify causes in data exploration. Export all tables and create visual reports to take into S&OP meetings and communicate forecasts and expected accuracy.
Forecast using Algorithms from Artificial Intelligence:
Artificial Neural Networks
Support Vector Regression
CART Decision Trees
Extreme Learning Machines
Algorithms from Statistics
Single Exponential Smoothing
Trend Exponential Smoothing
Seasonal Exponential Smoothing
Double Seasonal Smoothing
Triple Seasonal Smoothing
Croston's Exponential Smoothing
ARIMA (Box & Jenkins) Methods
Multiple Linear Regression
Seasonal Linear Regression
The latest collection of algorithms from computational intelligence, machine learning and data science for forecasting.
The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.