BEGIN:VCALENDAR VERSION:2.0 PRODID:-//ISOM 835 Fall 2026//isom-835.vercel.app//EN CALSCALE:GREGORIAN METHOD:PUBLISH X-WR-CALNAME:ISOM 835 — Predictive Analytics & ML (Fall 2026) X-WR-TIMEZONE:America/New_York BEGIN:VEVENT UID:session-1@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20260914T210000Z DTEND:20260914T233000Z SUMMARY:ISOM 835 · Session 1: From Data to Decisions: Your First Predict ive Model DESCRIPTION:What predictive analytics is\, why every business question a bout the future is a prediction problem — and a churn model you build y ourself before 7:30\nhttps://isom-835.vercel.app/sessions/session-01/ URL:https://isom-835.vercel.app/sessions/session-01/ END:VEVENT BEGIN:VEVENT UID:session-2@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20260921T210000Z DTEND:20260921T233000Z SUMMARY:ISOM 835 · Session 2: Data Wrangling & EDA for Prediction DESCRIPTION:Real data arrives in eight tables\, three date formats\, and a thousand blanks. pandas joins\, groupbys\, and datetime features on Olist e-commerce — and how to spot the leak that makes a model look bri lliant and ship useless\nhttps://isom-835.vercel.app/sessions/session-0 2/ URL:https://isom-835.vercel.app/sessions/session-02/ END:VEVENT BEGIN:VEVENT UID:session-3@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20260928T210000Z DTEND:20260928T233000Z SUMMARY:ISOM 835 · Session 3: Feature Engineering & Leak-Proof Pipelines DESCRIPTION:Encoders\, scalers\, imputers\, and the scikit-learn Pipelin e — the machinery that turns a messy table into a model input without e ver peeking at the test set\nhttps://isom-835.vercel.app/sessions/sessi on-03/ URL:https://isom-835.vercel.app/sessions/session-03/ END:VEVENT BEGIN:VEVENT UID:session-4@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261005T210000Z DTEND:20261005T233000Z SUMMARY:ISOM 835 · Session 4: Classification: Predicting Yes/No DESCRIPTION:Logistic regression as odds\, kNN and Naive Bayes as contras ts\, the confusion matrix in depth\, precision vs. recall\, ROC vs. PR — on 30\,000 credit-card customers\, 22% of whom default\nhttps://isom- 835.vercel.app/sessions/session-04/ URL:https://isom-835.vercel.app/sessions/session-04/ END:VEVENT BEGIN:VEVENT UID:session-5@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261019T210000Z DTEND:20261019T233000Z SUMMARY:ISOM 835 · Session 5: From Scores to Decisions: Thresholds\, Cos ts & Calibration DESCRIPTION:A model produces a probability\; a business produces a decis ion. Expected value\, profit curves\, tuned thresholds\, calibration\, lift and gain — and the launch of the Midterm Model Competition\nhttps: //isom-835.vercel.app/sessions/session-05/ URL:https://isom-835.vercel.app/sessions/session-05/ END:VEVENT BEGIN:VEVENT UID:session-6@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261026T210000Z DTEND:20261026T233000Z SUMMARY:ISOM 835 · Session 6: Regression: Predicting Numbers DESCRIPTION:What is a house worth? Linear regression as the interpretabl e baseline\, coefficients in dollars\, log targets\, ridge and lasso\, and the residual plot that tells you what the model is missing\nhttps:/ /isom-835.vercel.app/sessions/session-06/ URL:https://isom-835.vercel.app/sessions/session-06/ END:VEVENT BEGIN:VEVENT UID:session-7@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261102T220000Z DTEND:20261103T003000Z SUMMARY:ISOM 835 · Session 7: Decision Trees & Random Forests DESCRIPTION:How a tree asks questions\, why it memorizes\, and how avera ging hundreds of decorrelated trees turns a fragile model into the most reliable workhorse in tabular ML\nhttps://isom-835.vercel.app/sessions /session-07/ URL:https://isom-835.vercel.app/sessions/session-07/ END:VEVENT BEGIN:VEVENT UID:session-8@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261109T220000Z DTEND:20261110T003000Z SUMMARY:ISOM 835 · Session 8: Gradient Boosting & the Tabular Frontier DESCRIPTION:Fit the residuals\, again and again: HistGradientBoosting\, LightGBM\, XGBoost\, early stopping — and the tabular foundation models that\, since 2025\, beat tuned boosting on small data with zero tuning . Plus: the competition reveal\nhttps://isom-835.vercel.app/sessions/se ssion-08/ URL:https://isom-835.vercel.app/sessions/session-08/ END:VEVENT BEGIN:VEVENT UID:session-9@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261116T220000Z DTEND:20261117T003000Z SUMMARY:ISOM 835 · Session 9: Trust\, but Verify: Tuning\, Explaining & Uncertainty DESCRIPTION:Cross-validation that doesn't lie\, hyperparameter search wi th Optuna\, SHAP for one customer and the whole book\, conformal interv als that keep their promise\, fairness slices\, and the one-page model card\nhttps://isom-835.vercel.app/sessions/session-09/ URL:https://isom-835.vercel.app/sessions/session-09/ END:VEVENT BEGIN:VEVENT UID:session-10@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261123T220000Z DTEND:20261124T003000Z SUMMARY:ISOM 835 · Session 10: Segments\, Structure & Anomalies DESCRIPTION:Learning without labels: RFM customer segmentation with k-me ans\, choosing k honestly\, hierarchical clustering and DBSCAN\, PCA an d UMAP for seeing high-dimensional data\, and Isolation Forest for the transactions that don't belong\nhttps://isom-835.vercel.app/sessions/se ssion-10/ URL:https://isom-835.vercel.app/sessions/session-10/ END:VEVENT BEGIN:VEVENT UID:session-11@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261130T220000Z DTEND:20261201T003000Z SUMMARY:ISOM 835 · Session 11: Forecasting: Time Series With ML & Founda tion Models DESCRIPTION:Why random splits lie when time matters\, lag and rolling fe atures that turn forecasting into regression\, classical baselines you must beat\, and the zero-shot foundation models that changed the field in 2025\nhttps://isom-835.vercel.app/sessions/session-11/ URL:https://isom-835.vercel.app/sessions/session-11/ END:VEVENT BEGIN:VEVENT UID:session-12@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261207T220000Z DTEND:20261208T003000Z SUMMARY:ISOM 835 · Session 12: Neural Networks\, Embeddings & LLMs as Fe atures DESCRIPTION:A multilayer perceptron on tabular data (and why it usually loses to boosting)\, text as features — TF-IDF\, sentence embeddings\, zero-shot LLM labels — and the AI-assisted analyst's workflow: generate \, verify\, own\nhttps://isom-835.vercel.app/sessions/session-12/ URL:https://isom-835.vercel.app/sessions/session-12/ END:VEVENT BEGIN:VEVENT UID:session-13@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261214T220000Z DTEND:20261215T003000Z SUMMARY:ISOM 835 · Session 13: Ship It — Deployment\, Monitoring & Your Prediction Story DESCRIPTION:Persist the pipeline\, track it\, put a front end on it\, wa tch it drift\, know the rules — then the night the course was built for : your prediction stories\, presented\nhttps://isom-835.vercel.app/sess ions/session-13/ URL:https://isom-835.vercel.app/sessions/session-13/ END:VEVENT BEGIN:VEVENT UID:due-hw1@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261006T035959Z DTEND:20261006T035959Z SUMMARY:DUE · HW1 — EDA + baseline on Telco · Lab 0 Python Foundations C heck DESCRIPTION:Submit: notebooks (.ipynb). Points: 100 · completion.\nCanva s: https://canvas.suffolk.edu/courses/18389/assignments/259280\nAssigne d at https://isom-835.vercel.app/sessions/session-02#homework URL:https://canvas.suffolk.edu/courses/18389/assignments/259280 END:VEVENT BEGIN:VEVENT UID:due-hw2@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261020T035959Z DTEND:20261020T035959Z SUMMARY:DUE · HW2 — Credit default: classification\, ROC/PR\, odds ratio s DESCRIPTION:Submit: notebook (.ipynb). Points: 100.\nCanvas: https://can vas.suffolk.edu/courses/18389/assignments/261934\nAssigned at https://i som-835.vercel.app/sessions/session-04#homework URL:https://canvas.suffolk.edu/courses/18389/assignments/261934 END:VEVENT BEGIN:VEVENT UID:due-m1@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261103T045959Z DTEND:20261103T045959Z SUMMARY:DUE · Project Milestone 1 — your problem and your dataset DESCRIPTION:Submit: one paragraph (discussion post). Points: 5 · pass/fa il.\nCanvas: https://canvas.suffolk.edu/courses/18389/discussion_topics /202473\nAssigned at https://isom-835.vercel.app/project#milestones URL:https://canvas.suffolk.edu/courses/18389/discussion_topics/202473 END:VEVENT BEGIN:VEVENT UID:due-mid-pred@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261109T213000Z DTEND:20261109T213000Z SUMMARY:DUE · Midterm competition closes — predictions file DESCRIPTION:Submit: predictions (.csv)\, on Kaggle and on Canvas. Points : 50.\nCanvas: https://canvas.suffolk.edu/courses/18389/assignments/261 936\nAssigned at https://isom-835.vercel.app/sessions/session-05#homewo rk URL:https://canvas.suffolk.edu/courses/18389/assignments/261936 END:VEVENT BEGIN:VEVENT UID:due-mid-memo@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261110T045959Z DTEND:20261110T045959Z SUMMARY:DUE · Midterm competition memo DESCRIPTION:Submit: one page (.pdf). Points: 50.\nCanvas: https://canvas .suffolk.edu/courses/18389/assignments/261937\nAssigned at https://isom -835.vercel.app/sessions/session-08#homework URL:https://canvas.suffolk.edu/courses/18389/assignments/261937 END:VEVENT BEGIN:VEVENT UID:due-hw3@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261117T045959Z DTEND:20261117T045959Z SUMMARY:DUE · HW3 — Threshold & cost on the credit model + Ames regressi on DESCRIPTION:Submit: notebook (.ipynb). Points: 100.\nCanvas: https://can vas.suffolk.edu/courses/18389/assignments/261935\nAssigned at https://i som-835.vercel.app/sessions/session-06#homework URL:https://canvas.suffolk.edu/courses/18389/assignments/261935 END:VEVENT BEGIN:VEVENT UID:due-m2@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261124T045959Z DTEND:20261124T045959Z SUMMARY:DUE · Project Milestone 2 — baseline model with a metric DESCRIPTION:Submit: notebook (.ipynb). Points: 5 · pass/fail.\nCanvas: h ttps://canvas.suffolk.edu/courses/18389/assignments/261939\nAssigned at https://isom-835.vercel.app/project#milestones URL:https://canvas.suffolk.edu/courses/18389/assignments/261939 END:VEVENT BEGIN:VEVENT UID:due-m3@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261201T045959Z DTEND:20261201T045959Z SUMMARY:DUE · Project Milestone 3 — tuned model with an error analysis DESCRIPTION:Submit: notebook (.ipynb). Points: 5 · pass/fail.\nCanvas: h ttps://canvas.suffolk.edu/courses/18389/assignments/261940\nAssigned at https://isom-835.vercel.app/project#milestones URL:https://canvas.suffolk.edu/courses/18389/assignments/261940 END:VEVENT BEGIN:VEVENT UID:due-m4@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261207T220000Z DTEND:20261207T220000Z SUMMARY:DUE · Project Milestone 4 — dry-run slides\, then paired dry run s in class DESCRIPTION:Submit: slides (.pdf). Points: 5 · pass/fail.\nCanvas: https ://canvas.suffolk.edu/courses/18389/assignments/261941\nAssigned at htt ps://isom-835.vercel.app/project#milestones URL:https://canvas.suffolk.edu/courses/18389/assignments/261941 END:VEVENT BEGIN:VEVENT UID:due-final@isom-835.vercel.app DTSTAMP:20261014T220000Z DTSTART:20261214T220000Z DTEND:20261214T220000Z SUMMARY:DUE · Final project — notebook\, 2-page memo\, repo link · prese ntations that evening DESCRIPTION:Submit: notebook (.ipynb)\, memo (.pdf)\, repository URL. Po ints: 80 + 20.\nCanvas: https://canvas.suffolk.edu/courses/18389/assign ments/261942\nAssigned at https://isom-835.vercel.app/project URL:https://canvas.suffolk.edu/courses/18389/assignments/261942 END:VEVENT END:VCALENDAR