Table 3.

Rank of predictor and significant causality results for Paris.

Rank of Predictor	Number of repetitions	Percentage (%)	AC	AUPRC
PM2.5 → PM10	17,985	0.89	4.949	False
PM2.5 ← PM10	17,549	0.89	4.784	False
lnPM2.5 → lnPM10	18,161	0.89	4.745	False
lnPM2.5 ← lnPM10	18,542	0.89	4.197	False
dPM2.5 → dPM10	18,665	0.89	4.187	False
dPM2.5 ← dPM10	17,464	0.89	4.122	False
sPM2.5 → sPM10	17,952	0.89	4.125	False
sPM2.5 ← sPM10	17,465	0.89	4.896	False
d.lnPM2.5 → d.lnPM10	17,651	0.89	4.191	False
d.lnPM2.5 ← d.lnPM10	19,455	0.89	4.965	False
PM2.5 → Deaths	21,756	0.89	4.100	True
PM2.5 ← Deaths	21,479	0.89	4.105	False
lnPM2.5 → lnDeaths	16,984	0.89	4.204	False
lnPM2.5 ← lnDeaths	15,665	0.89	4.255	False
dPM2.5 → dDeaths	17,854	0.89	4.125	True
dPM2.5 ← dDeaths	16,249	0.89	4.202	False
sPM2.5 → sDeaths	19,845	0.89	4.265	False
sPM2.5 ← sDeaths	19,454	0.89	4.122	False
d.lnPM2.5 → d.lnDeaths	17,446	0.89	4.122	False
d.lnPM2.5 ← d.lnDeaths	17,445	0.89	4.122	False
PM10 → Deaths	17,945	0.89	4.365	True
PM10 ← Deaths	18,654	0.89	4.125	False
lnPM10 → lnDeaths	18,465	0.89	4.125	False
lnPM10 ← lnDeaths	19,542	0.89	4.136	False
dPM10 → dDeaths	19,451	0.89	4.795	True
dPM10 ← dDeaths	18,544	0.89	4.862	False
sPM10 → sDeaths	19,456	0.89	4.264	False
sPM10 ← sDeaths	21,949	0.89	4.166	False
d.lnPM10 → d.lnDeaths	20,495	0.89	4.102	False
d.lnPM10 ← d.lnDeaths	20,948	0.89	4.105	False

Notes: AC: Average Causality value; AUPRC: Area Under the Precision Recall Curve. True: P-Value < 0.05. False: P-Value ≥ 0.05. Number of repetitions: number of retries (0 and 1) carried out by the machine. Percentage (%): expected success rate compared to an opposite event.