Predict which travel modes a person uses, from their profile

Given only a person's age band, education, occupation, income group, children, vehicle ownership and transit pass, give the probability that they use active transport, public transport and a car.

Results in the task’s own metric

Klang Valley, Malaysia · Brier (sum over modes) ↓ · 200 items · Constant baseline: 0.524

Recomputed from saved answers. Model labels retain the original version and repeat protocol. No overall model average.

ModelRaw resultValid / asked
DeepSeek-V4-Flash (API, temp 0)0.665200 / 200
Llama-3.3-70B (API, temp 0)0.677200 / 200
Qwen3.8-27B (API, temp 0)0.682200 / 200
GLM-5.3 (API, temp 0)0.684200 / 200
MiniMax-M3 (API, temp 0)0.685200 / 200
gpt-oss-120B (API, temp 0)0.734200 / 200
Gemma-4-26B (API, temp 0)0.748200 / 200

Download current raw metrics (JSON)

Correct answer known from: The modes 200 residents of the Klang Valley reported using for daily short trips. The survey records mode use, not a single choice, so a person can be counted under two modes and the three probabilities need not sum to 1.

The sample's own usage rates score 0: active transport 64 per cent, car 74, public transport 11.