The Improvement Of Concept Lattice Construction Algorithm

The problem of concept lattice construction originated from philosophy.Nowadays,it has been widely used in practical production.It abstracts the common essential characteristics of things,summarizes them,forms the concept,and finally establishes a lattice structure in which the concepts are related to each other.That's the problem that the algorithm needs to solve.The purpose of this paper is to establish a concept lattice construction algorithm that implements double compression in time and space compared to the current algorithm.Firstly,according to the batch type,gradual structure type,parallel construction type and fuzzy structure type,the current stage construction algorithm is classified.In each class,several representative algorithms are selected to analyze the advantages and disadvantages,and finally select batch processing.The Chein algorithm in the model is the basis of our improved algorithm.The Chein algorithm has the advantage of not relying on the formal context,and the construction of the new concept and the hierarchical relationship between the concepts are independent of each other.Therefore,a method of dividing the subform background is proposed to divide the subform background into the new concept generation process according to the left-hand value of the intersection operator.The proof is given in the text.Therefore,independent calculations can achieve parallelization.After extensive experiments,it has been proved that using FP-Tree to store the concept can not only achieve a higher proportion of storage space compression,but also greatly optimize the inter-concept cross-computation process.The effect is obviously better than that of other algorithms.Another important improvement of this paper is to use FP-Tree to realize the early detection of the concept of redundancy and to directly delete the concept of redundancy.This avoids the generation of redundant concepts from the background division of the sub-situation and thus solves the waste of space caused by the redundant concept.A detailed proof was given.Finally,the multi-dimensional comparison between the standard UCI dataset and other algorithms in the field proves the correctness of the improved algorithm and the feasibility of high-performance.At the same time,the transposed data shows that the batch-type algorithm is compared with the progressive structure type.The algorithm has better practicality.The text also gives experimental data to facilitate other algorithmic analogies.