IDSC

Showing 25 of 37 from the live probe.

• Foretell Profile Classification

  Classify a time-series dataset for Foretell profiling. Provide the documented JSON request body as payload. In a full Foretell workflow, call change-point detection first and pass its inter_order_cpi so classification can report the latest segment as well as the whole series. Use final_characteristic when it is not NA; otherwise fall back to overall_characteristic. Universal user-approval rule: execute at most one business workflow stage per user approval. Calls needed only to finish that same s

• Foretell Detect Change Points

  Detect change points in a time-series dataset. Provide the documented JSON request body as payload. Decision rule for multiple named series: use foretell_detect_change_points_per_series by default when the user wants a result for each SKU, product, item, or series independently. Use foretell_detect_multivariate_change_points only when the user explicitly wants a common change point or a change in the relationship/correlation among aligned, equal-length series. Dictionary input is multivariate an

• Foretell Assess Predictability

  Assess time-series predictability. Provide the documented JSON request body as payload. In a full Foretell workflow, call this after change-point detection and pass its inter_order_cpi and order_quantity_cpi to score the detected regimes. Treat NA as insufficient evidence, not as a score. Universal user-approval rule: execute at most one business workflow stage per user approval. Calls needed only to finish that same stage, such as polling its asynchronous request, may continue without another a

• Foretell Profile Lost Sales Prophet

  Profile lost sales with Prophet. Provide the documented JSON request body as payload. Conditional profiling step: call only when stock availability constrained observed sales and lost-sales estimation is relevant. Use this Prophet variant for a continuous-characteristic series, pass target and stock_availability, and pass order_quantity_cpi when available. Do not infer that every zero sale is lost demand. Universal user-approval rule: execute at most one business workflow stage per user approval

• Foretell Profile Lost Sales Zinbtf

  Profile lost sales with ZINB/TensorFlow. Provide the documented JSON request body as payload. Conditional profiling step: call only when stock availability constrained observed sales and lost-sales estimation is relevant. Use this ZINB/TensorFlow variant for an intermittent-characteristic series, pass target and stock_availability, and pass order_quantity_cpi when available. Do not infer that every zero sale is lost demand. Universal user-approval rule: execute at most one business workflow stag

• Foretell Detect Anomalies

  Calculate time-series anomalies. Provide the documented JSON request body as payload. Optional diagnostic only: do not call anomaly detection by default in profiling or forecasting. Call it when the user explicitly asks about anomalies/outliers, when data quality is in question, or when unusual spikes or drops need investigation. A structural regime change belongs to change-point detection. Never remove or alter detected anomalies automatically without user confirmation or an explicit business r

• Foretell Select Features

  Select predictive features from training and prediction data. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: this dataframe-based ML workflow does not accept change-point-index fields and does not automatically discard earlier epochs. When change points were detected, restrict the training rows to the chosen latest relevant epoch before constructing training_dataset. Keep prediction_dataset unchanged and retain enough latest-epoch rows for the model to train.

• Foretell Forecast Continuous Prophet

  Forecast a continuous time series with Prophet. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: after change-point detection, pass the full original time_series plus the detected change-point indices to this Prophet request. For continuous demand, pass order_quantity_cpi. For intermittent demand, pass both order_quantity_cpi and inter_order_cpi. The Foretell backend already restricts training to the latest relevant epoch starting at the last supplied change po

• Foretell Train Continuous Prophet

  Train a continuous Prophet model. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: after change-point detection, pass the full original time_series plus the detected change-point indices to this Prophet request. For continuous demand, pass order_quantity_cpi. For intermittent demand, pass both order_quantity_cpi and inter_order_cpi. The Foretell backend already restricts training to the latest relevant epoch starting at the last supplied change point. Do not pr

• Foretell Predict Continuous Prophet

  Predict with a trained continuous Prophet model. Provide the documented JSON request body as payload. Universal user-approval rule: execute at most one business workflow stage per user approval. Calls needed only to finish that same stage, such as polling its asynchronous request, may continue without another approval. After the stage reaches a final result, summarize the outcome, identify the recommended next stage and its required inputs, then stop. Never call an API for the next business stag

• Foretell Forecast Intermittent Prophet

  Forecast an intermittent time series with Prophet. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: after change-point detection, pass the full original time_series plus the detected change-point indices to this Prophet request. For continuous demand, pass order_quantity_cpi. For intermittent demand, pass both order_quantity_cpi and inter_order_cpi. The Foretell backend already restricts training to the latest relevant epoch starting at the last supplied change

• Foretell Train Intermittent Prophet

  Train an intermittent Prophet model. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: after change-point detection, pass the full original time_series plus the detected change-point indices to this Prophet request. For continuous demand, pass order_quantity_cpi. For intermittent demand, pass both order_quantity_cpi and inter_order_cpi. The Foretell backend already restricts training to the latest relevant epoch starting at the last supplied change point. Do not

• Foretell Predict Intermittent Prophet

  Predict with a trained intermittent Prophet model. Provide the documented JSON request body as payload. Universal user-approval rule: execute at most one business workflow stage per user approval. Calls needed only to finish that same stage, such as polling its asynchronous request, may continue without another approval. After the stage reaches a final result, summarize the outcome, identify the recommended next stage and its required inputs, then stop. Never call an API for the next business st

• Foretell Predict Continuous Xgboost

  Predict a continuous series with XGBoost. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: this dataframe-based ML workflow does not accept change-point-index fields and does not automatically discard earlier epochs. When change points were detected, restrict the training rows to the chosen latest relevant epoch before constructing training_dataset. Keep prediction_dataset unchanged and retain enough latest-epoch rows for the model to train. Universal user-appr

• Foretell Predict Continuous Explanatory Prophet

  Predict a continuous series with explanatory Prophet. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: this dataframe-based ML workflow does not accept change-point-index fields and does not automatically discard earlier epochs. When change points were detected, restrict the training rows to the chosen latest relevant epoch before constructing training_dataset. Keep prediction_dataset unchanged and retain enough latest-epoch rows for the model to train. Univers

• Foretell Predict Continuous Lightgbm

  Predict a continuous series with LightGBM. Provide the documented JSON request body as payload. Latest-epoch forecasting rule: this dataframe-based ML workflow does not accept change-point-index fields and does not automatically discard earlier epochs. When change points were detected, restrict the training rows to the chosen latest relevant epoch before constructing training_dataset. Keep prediction_dataset unchanged and retain enough latest-epoch rows for the model to train. Universal user-app

• Foretell Detect Change Points Per Series

  Detect change points independently for each SKU or named series. Use this by default when the user supplies multiple SKUs and wants a result for each one. This is one MCP call but one IDSC API request per series, so each successful series is processed and charged independently. Do not use this for relationship or correlation regime changes across series.

• Foretell Detect Multivariate Change Points

  Detect a common relational change point across aligned series. Use only when the user explicitly asks about a shared regime change or a change in relationship/correlation among at least two series. All series must represent the same ordered periods and have equal lengths. This returns one multivariate result, not separate results per SKU.

• Foretell Profile Demand

  Run the complete core Foretell profiling stage in one approved MCP call. This calls change-point detection, classification, and predictability in sequence. Lost-sales profiling is added only when stock availability was supplied, shows constrained sales, and the classification maps safely to a documented lost-sales model. It does not run anomaly detection or start a forecast. Each successful upstream call may consume API credit.

• Optimax Product Mix

  Run Optimax product-mix optimization. Provide the documented JSON request body as payload. Operational planning stage after tactical Category Mix in a complete Optimax workflow. Product Mix optimizes SKUs at one location and does not understand category constraints. It can solve SKUs from multiple categories together when they intentionally share one space_capacity. When Category Mix has reserved capacity per category, prefer a separate Product Mix call for each location and category, include on

• Optimax Category Mix

  Run Optimax category-mix optimization. Provide the documented JSON request body as payload. Tactical allocation stage after S&OP and the paired stage before operational Product Mix in a complete Optimax workflow. Use Category Mix when the user must allocate available capacity across categories or locations. Its category_details result gives the planned quantity per category at each location; derive that category's prepared space as allocated quantity multiplied by category_data[category].space_p

• Mcf Start Optimization

  Start asynchronous multi-commodity-flow optimization. Provide the documented JSON request body as payload. Strategic planning stage and first stage of a complete Optimax planning workflow. Use MCF to decide the long-term supply-chain network, node activation, production transformation, capacity, and commodity flows. Because it is asynchronous, obtain its FINISH result before proceeding to downstream tactical planning. Universal user-approval rule: execute at most one business workflow stage per

• Sop Start Optimization

  Start asynchronous S&OP optimization. Provide the documented JSON request body as payload. Tactical planning stage after strategic MCF in a complete Optimax workflow. Use S&OP for time-phased production, procurement, inventory, manpower, and demand-fulfilment planning within the strategic network assumptions. Because it is asynchronous, obtain its FINISH result before downstream category allocation. Universal user-approval rule: execute at most one business workflow stage per user approval. Call

• Optfleet Start Static Routing

  Start asynchronous static fleet routing. Provide the documented JSON request body as payload. Optfleet is a separate fleet-routing/load-planning workflow. Do not call it automatically after Foretell or an Optimax planning stage; use it only when the user explicitly asks for transport routing, live route changes, or vehicle loading and supplies the required logistics data. Universal user-approval rule: execute at most one business workflow stage per user approval. Calls needed only to finish that

• Optfleet Start Dynamic Job Ica

  Start asynchronous dynamic fleet job insertion, cancellation, or adjustment. Provide the documented JSON request body as payload. Optfleet is a separate fleet-routing/load-planning workflow. Do not call it automatically after Foretell or an Optimax planning stage; use it only when the user explicitly asks for transport routing, live route changes, or vehicle loading and supplies the required logistics data. Universal user-approval rule: execute at most one business workflow stage per user approv