Table 2 A comparison between model parameters \(\beta\) learned using the proposed protocols and ordinary logistic regression protocol over LBW dataset
From: Privacy-preserving logistic regression with secret sharing
\(\beta\) | Ordinary LR | Accurate BMPC | BMPC | Accurate CMPC | CMPC | ||||
|---|---|---|---|---|---|---|---|---|---|
3 | 5 | 7 | 3 | 5 | 7 | ||||
\(\beta _1\) | 0.01574 | 0.01577 | 0.01761 | 0.01580 | 0.01480 | 0.01574 | 0.02214 | 0.01793 | 0.01630 |
\(\beta _2\) | 0.01127 | 0.01123 | 0.01171 | 0.01061 | 0.01006 | 0.01127 | 0.01534 | 0.01266 | 0.01166 |
\(\beta _3\) | 0.78666 | 0.78152 | 0.67763 | 0.62479 | 0.60392 | 0.78662 | 0.95081 | 0.81191 | 0.77010 |
\(\beta _4\) | \(-0.47132\) | \(-0.46992\) | \(-0.48975\) | \(-0.44423\) | \(-0.42170\) | \(-0.47131\) | \(-0.63960\) | \(-0.52621\) | \(-0.48340\) |
\(\beta _5\) | \(-1.32410\) | \(-1.31870\) | \(-1.24442\) | \(-1.13786\) | \(-1.08974\) | \(-1.32405\) | \(-1.68676\) | \(-1.41595\) | \(-1.32408\) |
\(\beta _6\) | \(-0.75584\) | \(-0.75594\) | \(-0.86971\) | \(-0.78142\) | \(-0.73330\) | \(-0.75583\) | \(-1.09894\) | \(-0.88944\) | \(-0.80596\) |
\(\beta _7\) | \(-2.20748\) | \(-2.20104\) | \(-2.48191\) | \(-2.23117\) | \(-2.09511\) | \(-2.20743\) | \(-3.15262\) | \(-2.56252\) | \(-2.33208\) |
\(\beta _8\) | \(-0.96060\) | \(-0.95756\) | \(-0.99906\) | \(-0.90459\) | \(-0.85667\) | \(-0.96058\) | \(-1.30317\) | \(-1.07358\) | \(-0.98838\) |
\(\beta _9\) | \(-0.24569\) | \(-0.24476\) | \(-0.21884\) | \(-0.20160\) | \(-0.19465\) | \(-0.24568\) | \(-0.30509\) | \(-0.25879\) | \(-0.24367\) |