Skip to main content

Documentation Index

Fetch the complete documentation index at: https://cors-lau.vercel.app/docs/llms.txt

Use this file to discover all available pages before exploring further.

This page collects answers to the most frequently asked questions about CORS. If your question is not covered here, refer to the relevant guide or use the feedback link at the bottom of the page.
If a course has no historical offerings records, the Recommendation Engine cannot calculate an enrollment-based latent demand estimate for it. Instead, CORS falls back to the enrollment influx numbers you enter before running the analysis — the expected counts for new freshmen, new sophomores, and new masters students. These are used as a proxy for potential demand.Prediction quality for courses without historical data will be lower than for well-documented courses. After the course is offered and outcomes are recorded, upload the updated historical data and click Recalibrate Intelligence to incorporate it into future analyses.
CORS currently supports two campuses: Beirut and Byblos. These campuses are fixed in the current version of the platform. When running an analysis or scheduling a timetable, select the campus that matches the data you are working with.If your institution adds a new campus or a new site that requires separate analysis, contact your CORS administrator or refer to the deployment documentation for configuration options.
Yes, but you run them sequentially rather than simultaneously. CORS maintains a separate ensemble ML model for each campus. To generate recommendations for both Beirut and Byblos:
  1. Open the Recommendation Engine, select Beirut, configure your settings, and click Execute Analysis.
  2. Once complete, return to the Recommendation Engine, select Byblos, and run the analysis again.
Each run produces an independent results set for that campus.
Recalibrate Intelligence retrains the ensemble ML models using all current historical offerings data on record for the selected campus. The process updates the model weights so that new enrollment outcomes and newly uploaded semesters are reflected in future analyses.You should recalibrate after:
  • Uploading a new batch of historical offerings data
  • Adding or updating your course catalog
  • Receiving significantly different enrollment numbers than past semesters
Recalibration does not affect in-progress analyses. Start a new analysis after recalibrating to use the updated models.
The Model ROC-AUC metric on the Dashboard is calculated from the most recent prediction run. If no analysis has been executed yet — or if the model has been recalibrated but no new analysis has been run since — the metric shows N/A because there are no prediction results to evaluate against.To populate this metric, run an analysis from the Recommendation Engine. The ROC-AUC value will update automatically once predictions are available.
Strict Capacity Quotas mode overrides the standard ML probability threshold. Instead of marking all courses above a certain confidence level as Recommended, the engine selects exactly the top N courses per department-and-type slot, ranked by offer score.This is useful when you have firm staffing or room constraints and cannot accommodate an open-ended list of recommendations. You configure the slot limits (e.g. CSC Core: 12, MTH: 8) before running the analysis.See Run your first course recommendation analysis for default slot values and configuration instructions.
Projected sections is derived from latent demand divided by the typical section capacity for the course level:
  • 100- to 300-level courses use a standard section capacity based on historical averages.
  • 400-level courses receive a downward adjustment because senior-level courses typically run with smaller cohorts.
The result is rounded up to ensure demand can be met. Treat the projected section count as a recommendation — you may increase or decrease it based on available rooms and faculty.
Yes. In the Timetable Orchestration view, click any scheduled block on the timetable grid to open the detail drawer. The drawer shows the current assignment details and includes a Modify Assignment button.Use Modify Assignment to change the assigned professor or room without moving the time slot. To change the day pattern or start time, remove the placement and re-place the course in the new slot.After making changes, click Finalize & Export again to download an updated cors_final_schedule.xlsx with your revisions.
CORS does not enforce a documented maximum file size for data uploads. However, very large files — particularly historical offerings spreadsheets spanning many years and thousands of rows — may slow down the ingestion and recalibration process.For best performance, keep individual upload files to a reasonable size. If you have a very large historical dataset, consider splitting it by academic year before uploading, then uploading each file separately.
A De-prioritized status means the ML model assigned the course an offer score below the recommendation threshold. In practical terms, the model predicts that demand for the course is low enough, and its impact on degree progression modest enough, that it does not need to be offered in the upcoming semester.De-prioritized does not mean the course is blocked. You can still add any de-prioritized course to your timetable manually in the Timetable Scheduler if you have programmatic, accreditation, or policy reasons to offer it. The model’s recommendation is advisory.
When you confirm a course placement in the Timetable Scheduler and a conflict exists, a red error banner describes the specific problem. The three types of conflicts and their resolutions are:
ConflictResolution
Professor double-bookingChoose a non-overlapping time slot, or assign a different professor
Room conflictSelect a different room for the same time slot
Professor availability violationMove the session to a time within the professor’s configured availability window, or assign a professor without that restriction
After making the adjustment, click Confirm again. CORS re-checks for conflicts before saving the placement.
Clicking Finalize & Export downloads cors_final_schedule.xlsx — a Microsoft Excel workbook. Each row represents one scheduled section and includes the following columns:
ColumnDescription
Days PatternDay code string (e.g. MWF, TR)
Course CodeCourse identifier (e.g. CSC201)
Section NameSection label
InstructorAssigned professor’s name
Start TimeSession start in HH:MM format
End TimeSession end in HH:MM format
RoomAssigned room name or number
The file can be opened directly in Excel, Google Sheets, or any application that supports the XLSX format.

​
Still have questions?

Guides

Step-by-step walkthroughs for common CORS workflows.

Data format reference

Detailed specifications for every file type CORS accepts.