摘要：AbstractProblem Statement. Among the main problems of e-Learning one is t he “cold st art problem”. A learning environment cannot provide informat ion on a relevant difficulty of t he cont ent due t o t he lack of informat ion about a learner. The unsolved problem ent ails a reduct ion in t he effect iveness of t he learning process as overly difficult or, on t he cont rary, t oo easy cont ent leads t o a loss of learning mot ivat ion, frust rat ion and st ress among st udent s. The st udy is aimed at searching t he solut ion of t he “cold st art problem”.Research Questions. Which model of det ect ing t he st art ing difficult y of t he cont ent will be universal and wil l produce a st able forecast for different samples?Purpose of the Study. Developing t he model for det ect ing t he opt imal st art ing difficult y of t he cont ent, including searching predict ors meet s the crit eria of t he universalit y. T est ing hypotheses about t he sust ainabilit y of t he model for samples wit h different levels of preparedness.Research Methods. The model developing based on Logist ic regression and It em Response T h eory. T est ing hypot heses about sust ainabilit y base d on hybrid simulat ion. This simulat ion type used real predict ors and generat ed dat a set s (paramet ers of t he cont ent difficult y) simult aneously. The dat a set s have made by Mont e-Carlo soft ware. Simulat ion has replicat ed several t imes for checking t he sust ainabilit y crit eria.Findings. Verbal and numerical int elligence paramet ers are pot ent and universal learning efficiency predict ors. T he det ect ed st art ing level of t he cont ent difficulty is close t o simulat ed learners’ level of knowledge (preparedness). The model is st at ist ical significant and sust ainable in cases of samples wit h similar or different levels of knowledge (preparedness).
关键词：e-Learning;cold start problem;verbal intelligence;nimerical intelligence;item response theory
