| Data mining refers to the process, which is to find the unknown and usefulknowledge (or module) from large amount of data. It is an important subjectafter the database artificial etc. For the mined knowledge can give powerfulsupport to its field, the data mining technology is widely applied. In theresearch of data mining algorithm, the more influential one is associationalrule finding algorithm. The associational algorithm is an important branchin the data mining research. It is also the most typical one in all kinds ofdata mining knowledge.At first, we discussed the function and the traditional methods of datamining, the emphasis is FP-growth algorithm which belongs to association rulealgorithm, and then, we compare the FP-growth algorithm and Apriori algorithmby doing experiment. After analyzing the FP-growth algorithm carefully, wemake improvement of FP-growth algorithm, we make improvement of FP-growthalgorithm on three points: 1. the bidirectional FP-tree is changed to be aunidirectional tree for predigest the data structure. 2. Conditional FP-treeshouldn't be created by means of change data structure. 3. In the process ofthe algorithm, loop is used to instead of recursion, which can lower thespending of the memory. The performance of improvement of FP-growth algorithmis proved to be better by analyzing the algorithm from timeand space, andthe experiment.Based on accomplishing system design and database design of the collegiatetechnological management system, we decided that association rule algorithmcan be used in it. After researching and improving the association rulealgorithm, we mine the six tables which is preprocessed by improved FP-growthalgorithm. After that, we conclude the college should pay more attention toapply for the projects in national level and more and more scholar shouldattend the project.In the last part of this dissertation, we summarized our work and analyzedthe improvements to be done in the future. |
PDF Full Text Request