Алгоритм мониторинга выпускников вузов на основе модели колебаний, таблиц сопряжённости и кластерного анализа методом K-средних
- № 3 (43) 2017
Страницы:
19
–
26
Язык: русский
Аннотация
В статье описана модель изменения количества выпускников по направлению профильной занятости, приведены таблицы сопряжённости, выбранный критерий, проведен кластерный анализ методом k-средних и дан общий алгоритм мониторинга на основе данных методов. Для выявления значимых индикаторов, влияющих на профильное трудоустройство выпускников, используются таблицы сопряженности и критерий χ2. Для классификации выпускников на основе выделенных индикаторов профильности их трудоустройства использован кластерный анализ методом k-средних.
Мақолада битирувчиларнинг сонини ўзгариши услуби – мутахассислик бўйича бандлик йўналишида, χ2 мезонлари, k-ўртача кластер таҳлил услубида ва шу услублар асосидаги умумий алгоритм мониторинги. Битирувчиларни касб бўйича ишга жойлаштириш учун аҳамиятли индикаторларни аниқланишда мослашиш жадвали ва χ2 мезонларидан фойдаланилади. Битирувчиларни классификацияси учун аниқланган индикаторлар мутахассислиги асосида, уларни ишга жойлаштиришда, k-ўртача услубидаги кластер таҳлилдин фойдаланилган.
The article describes a model of changing the number of graduates in the direction of the employment profile, contingency tables, the χ2 test, cluster analysis method of k-means and the overall monitoring algorithm, based on these methods. The essence of the model of changing the number of graduates from the profile employment is to determine the deviations from the equilibrium position between the profile employment of university graduates and the number of graduates in the direction of training, including data on the employment profile. Fluctuations model provides not only performance deviation of the number of graduates from the equilibrium value, but also a number of characteristics that influence on this process. In order to identify meaningful indicators that affect on the employment profile of graduates used contingency tables and χ2 test. When carrying out the chi-square test, the mutual independence of the two variables of the conjugacy table is checked, and due to this, the dependence of both variables indirectly explained. For conclusions about the relationship of direction were used ordinal scale and the correlation coefficients of Spearman. For the classification of the graduates on the basis of the selected profiled employment indicators, used k-means clustering. The essence of cluster analysis is the partitioning of a set of objects into groups, called clusters. Within each group should appear similar objects, and the objects of different groups should be as different as possible.
Мақолада битирувчиларнинг сонини ўзгариши услуби – мутахассислик бўйича бандлик йўналишида, χ2 мезонлари, k-ўртача кластер таҳлил услубида ва шу услублар асосидаги умумий алгоритм мониторинги. Битирувчиларни касб бўйича ишга жойлаштириш учун аҳамиятли индикаторларни аниқланишда мослашиш жадвали ва χ2 мезонларидан фойдаланилади. Битирувчиларни классификацияси учун аниқланган индикаторлар мутахассислиги асосида, уларни ишга жойлаштиришда, k-ўртача услубидаги кластер таҳлилдин фойдаланилган.
The article describes a model of changing the number of graduates in the direction of the employment profile, contingency tables, the χ2 test, cluster analysis method of k-means and the overall monitoring algorithm, based on these methods. The essence of the model of changing the number of graduates from the profile employment is to determine the deviations from the equilibrium position between the profile employment of university graduates and the number of graduates in the direction of training, including data on the employment profile. Fluctuations model provides not only performance deviation of the number of graduates from the equilibrium value, but also a number of characteristics that influence on this process. In order to identify meaningful indicators that affect on the employment profile of graduates used contingency tables and χ2 test. When carrying out the chi-square test, the mutual independence of the two variables of the conjugacy table is checked, and due to this, the dependence of both variables indirectly explained. For conclusions about the relationship of direction were used ordinal scale and the correlation coefficients of Spearman. For the classification of the graduates on the basis of the selected profiled employment indicators, used k-means clustering. The essence of cluster analysis is the partitioning of a set of objects into groups, called clusters. Within each group should appear similar objects, and the objects of different groups should be as different as possible.