High Performance Gibbs Sampling for IRT Models Using Row-Wise Decomposition

<table>Domain decomposition of the input <svg height="11.45" id="M54" style="vertical-align:-2.79825pt" version="1.1" viewbox="0 0 8.625 11.45" width="8.625" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<g transform="matrix(.017,-0,0,-.017,.062,7.9)"><path d="M330 461h152v-24q-24 -3 -36 -14t-24 -41l-148 -398q-45 -121 -77 -156q-31 -33 -84 -33q-43 0 -70 22t-27 57q0 29 17 48t45 19q25 0 42.5 -15t17.5 -37l-2 -14l-1 -6q0 -7 5 -12t12 -5q38 0 66 82l18 53l-119 290q-49 121 -67 146q-7 9 -34 14v24h249v-24l-16 -1
q-36 -3 -36 -25q0 -6 20 -58l72 -186l68 188q11 33 11 49q0 18 -11 25t-43 8v24z" id="x1D432"></path></g>
</svg> matrix and item parameter vectors by column using five processing nodes.</table>

International Scholarly Research Notices

fig1

Figure 1

Figure 1: High Performance Gibbs Sampling for IRT Models Using Row-Wise Decomposition 