BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//cfp.jupytercon.com//ACPULZ
BEGIN:VTIMEZONE
TZID:Europe/Paris
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-2023-ACPULZ@cfp.jupytercon.com
DTSTART;TZID=Europe/Paris:20230512T103000
DTEND;TZID=Europe/Paris:20230512T130000
DESCRIPTION:While the tutorial attendance is comprised in the conference p
ass\, we ask you to register for this tutorial on https://www.jupytercon.c
om/tickets as the seats available are limited.\n-- \n\nTutorial notebooks:
\n\n- https://github.com/soda-inria/survival-analysis-benchmark\n\nAccordi
ng to Wikipedia:\n\nSurvival analysis is a branch of statistics for analyz
ing the expected duration of time until one event occurs\, such as deaths
in biological organisms and failure in mechanical systems. [...]. Survival
analysis attempts to answer certain questions\, such as what is the propo
rtion of a population which will survive past a certain time? Of those tha
t survive\, at what rate will they die or fail? Can multiple causes of dea
th or failure be taken into account? How do particular circumstances or ch
aracteristics increase or decrease the probability of survival?\n\nIn this
two-part tutorial (morning and afternoon)\, we will deep dive into a prac
tical case study of predictive maintenance using tools from the scientific
Python ecosystem. Here is a tentative agenda:\n\nPart 1 (Morning)\n- What
is time-censored data and why it is a problem to train time-to-event regr
ession models.\n- Single event survival analysis with Kaplan-Meier using s
cikit-survival.\n- Competing risks modeling with Nelsonâ€“Aalen\, Aalen-Jo
hansen using lifelines.\n- Evaluation of the calibration of survival analy
sis estimators using the integrated brier score (IBS) metric.\n- Predictiv
e survival analysis modeling with Cox Proportional Hazards\, Survival Fore
sts using scikit-survival\, GradientBoostedIBS implemented from scratch wi
th scikit-learn.\n- Estimation of the cause-specific cumulative incidence
function (CIF) using our GradientBoostedIBS model.\n\nPart 2 (Afternoon)\n
- How to use a trained GradientBoostedIBS model to estimate the median sur
vival time and the probability of survival at a fixed time horizon.\n- Mea
suring the statistical association between input features and survival pro
babilities using partial dependence plot and permutation feature importanc
e.\n- Presentation of the results of a benchmark of various survival analy
sis estimators on the KKBox dataset.\n- Extracting implicit failure data f
rom operation logs using sessionization with Ibis and DuckDB.\n- Hands-on
wrap-up exercise.\n\nIt is not recommended to attend Part 2 without having
attended Part 1.\n\nTarget audience: good familiarity with machine learni
ng concepts\, with prior experience using scikit-learn (you know what cros
s-validation means and how to fit a Random Forest on a Pandas dataframe).
DTSTAMP:20230528T202246Z
LOCATION:Room 3 (Tutorial)
SUMMARY:Predictive survival analysis and competing risk modeling with sciki
t-learn\, scikit-survival\, lifelines\, Ibis\, and DuckDB (Part 1) - Olivi
er Grisel\, Guillaume Lemaitre\, Vincent Maladiere
URL:https://cfp.jupytercon.com/2023/talk/ACPULZ/
END:VEVENT
END:VCALENDAR