DescriptionCEESIt is a computational method for the analysis of short tandem repeats (STRs) in DNA for human identification. CEESIt computes the likelihood ratio (LR), the ratio of the probability of the evidence (the electropherogram obtained from the DNA sample) given a specific person of interest (POI) to the probability of the evidence given a random contributor from the background population. The DNA sample may be a mixture, comprised of multiple contributors at different ratios. With cases using low amounts of template DNA or cases with multiple contributors in the mixture, the results lacked consistency between computations. With 1-person mixtures, the tests ran with high repeatability and short runtime but with 2-people mixtures, the results had varying results and significantly longer runtime. The goal was to find the source of the discrepancies to improve repeatability and accuracy. CEESIt uses the Monte Carlo Method to generate the final probabilistic values. To improve repeatability and accuracy, importance sampling of the genotypes of the background population was implemented. By careful sampling and appropriate weighting to represent the background population, this improved the overall accuracy in the algorithm and allowed the algorithm to sample a smaller population, which decreases runtime.