Win­ter 2020/21: Baye­si­an Sta­tis­tics I

Con­t­ents

Baye­si­an thin­king dif­fers from fre­quen­tist sta­tis­tics in its in­ter­pre­ta­ti­on of pro­ba­bi­li­ty and un­cer­tain­ty. It com­ple­ments the exis­ting sta­tis­ti­cal tool­box with power­ful me­thods for si­mu­la­ti­on and in­fe­rence. The lec­tu­res Baye­si­an Sta­tis­tics I and II aim to fa­mi­lia­ri­ze the stu­dents to the Baye­si­an ap­proach. The first part deals with the theo­re­ti­cal fun­da­men­tals and the princi­ples of esti­ma­ting, tes­ting, fo­re­cas­ting and model as­sess­ment. In ad­di­ti­on, Baye­si­an re­gres­si­on con­cepts and computer-​​in­ten­si­ve si­mu­la­ti­on me­thods such as Mar­kov chain Monte Carlo (MCMC) are in­tro­du­ced. The se­cond part com­ple­ments and deepens these to­pics, for examp­le by Baye­si­an non­pa­ra­metric den­si­ty esti­ma­ti­on, Baye­si­an model choice and Ap­pro­xi­ma­te Baye­si­an Com­pu­ting (ABC).

Ge­ne­ral in­for­ma­ti­on

Lec­tu­rers: Prof. Dr. Chris­tia­ne Fuchs (lec­tu­res), Houda Ya­qi­ne (ex­er­ci­s­es)

Type: Lec­tu­re with (op­tio­nal but re­com­men­ded) ex­er­ci­s­es

Study achie­vements (Stu­di­en­leis­tun­gen): Study achie­vements for the ex­er­ci­se clas­ses can be ful­fil­led by pre­pa­ra­ti­on and one sub­mis­si­on of ex­er­ci­se sheets.

Re­com­men­ded pre­re­qui­si­tes: Good know­ledge of sta­tis­tics (esp. (con­di­tio­nal) den­si­ties/pro­ba­bi­li­ties, li­keli­hood in­fe­rence, re­gres­si­on) and R

Mo­du­le al­lo­ca­ti­on: see eKVV (lec­tu­re) and eKVV (ex­er­ci­s­es)

Dates: The lec­tu­res and ex­er­ci­se clas­ses take place in an on­line for­mat (live zoom con­fe­ren­ces, sin­gle lec­tu­res po­ten­ti­al­ly also as re­cor­ded vi­de­os). Live mee­tings will be on Thurs­days and Fri­days as lis­ten in the fol­lo­wing. Plea­se check this page re­gu­lar­ly for up­dates! Ac­cess de­tails are sent to the stu­dents via email / eKVV.

Li­te­ra­tu­re

• Lee: Baye­si­an Sta­tis­tics. Wiley, 4th edi­ti­on.
• Gel­man et al.: Baye­si­an Data Ana­ly­sis. CRC Press, 3rd edi­ti­on.
• Held & Sabanés Bové: Ap­p­lied Sta­tis­ti­cal In­fe­rence. Sprin­ger.

Ma­te­ri­al

Lec­tu­re sli­des, ex­er­ci­se sheets and fur­ther ma­te­ri­al will be made avail­able via Lern­raum­Plus.

This class is sup­por­ted by Data­Camp, a lear­ning plat­form for data sci­ence. Mem­bers of this class can ac­cess all cour­ses for free. The in­vi­ta­ti­on link is avail­able th­rough Lern­raum­Plus.