My next article is on fitting a Bradley-Terry model using logistic regression. This setting makes it very easy to adding additional factors (home/away/neutral, days between games, etc.).
You can see my code here - "uscho_krach+.R" is KRACH with home field factor, and the output is "uscho_krach+.txt".
https://github.com/octonion/hockey/t...ster/lunchtime
You can see my code here - "uscho_krach+.R" is KRACH with home field factor, and the output is "uscho_krach+.txt".
https://github.com/octonion/hockey/t...ster/lunchtime
Code:
Estimate Quinnipiac 3.95033463 Massachusetts-Lowell 3.25243187 Minnesota 2.82388531 Yale 2.27942112 Miami 2.21105503 Notre Dame 2.17199581 New Hampshire 2.13214626 Boston College 2.12837958 St. Cloud State 2.00911565 Union 2.00624919 North Dakota 1.99054169 Minnesota State 1.96780382 Denver 1.84893604 Wisconsin 1.82564064 Providence 1.70396231 Boston University 1.68374652 Western Michigan 1.66448139 Rensselaer 1.66198638 Brown 1.40413965 Colorado College 1.39121085 Cornell 1.38824557 Niagara 1.36282892 Dartmouth 1.35395617 St. Lawrence 1.26579460 Ohio State 1.20216401 Ferris State 1.19325555 Nebraska-Omaha 1.16140236 Alaska 1.12382232 Michigan 1.11264315 Merrimack 1.07745208 Northern Michigan 1.05049968 Robert Morris 0.95742202 Colgate 0.95659683 Bowling Green 0.95496607 Vermont 0.87329083 Holy Cross 0.84227822 Lake Superior 0.83855096 Maine 0.83670908 Massachusetts 0.83584215 Princeton 0.82634702 Connecticut 0.81964119 Minnesota-Duluth 0.81804466 Michigan Tech 0.80212041 Air Force 0.79966843 Michigan State 0.77855651 Harvard 0.75656645 Mercyhurst 0.75471428 Canisius 0.70330624 Clarkson 0.70023157 Penn State 0.64530158 RIT 0.58604616 Northeastern 0.52206478 Bemidji State 0.45855441 American International 0.41921680 Alaska-Anchorage 0.35397604 Bentley 0.34369988 Army 0.21655539 Alabama-Huntsville 0.07835112 Sacred Heart 0.07725705