Job Recommendation System Based on Analytic Hierarchy Process and K-means Clustering

Many students search for summer jobs during the vacation, but there are always too many choices. We need to find a way to help people choose a best summer job. We constructed a three-tier system to comprehensively illustrate the factors that high school students need to consider when looking for a summer job from the criteria of comfort, salary, personal gain, and matching degree. Under each criterion lie several sub-criteria (which are discussed later in detail). We also investigated students’ opinions toward each factor to get the judgement matrices for our AHP model. To reduce the subjectivity of the AHP model and reduce the correlation of various indexes in model construction, the AHP model and principal component analysis model were combined to construct the optimal weight model to obtain the optimal weight. And we utilized K-means clustering model to classify the work, adopted elbow method to determine the K value of the number of categories divided according to SSE (Sum of the squared errors) from the perspective of the data itself, and selected the class with the highest clustering center as the selection range of students. Finally we created ten fictional persons based on the samples we chose. The relevant questionnaires tested the students' character ability, and we used the GRNN neural network model to map the questionnaire to the weight. In this way, our model can conveniently get the weight result and calculate to help students find the optimal jobs collection by filling in the questionnaire.