The Methods Of Rule Extraction Based On 2-tuple Linguistic Formal Decision Context

In the era of big data,there are massive multi-type fuzzy data.In real life,people are more inclined to use natural language to describe and process such information.As an effective data analysis tool to describe the relationship between objects and attributes,concept lattice is an important means to extract rules,and linguistic-valued rule extraction has also become a hot issue.In order to reduce the information loss in fuzzy linguistic information processing,based on the 2-tuple linguistic model,this thesis proposes a 2-tuple linguistic normalized method of multi-type fuzzy data and the conversion method between 2-tuple linguistic formal context of different granularity.The rule extraction methods based on the2-tuple linguistic formal decision context are analyzed.The main research results are as follows:1.Aiming at the complexity of the initial fuzzy data types in the process of decision making and reasoning,the conversion models of fuzzy sets,intuitionistic fuzzy sets,hesitant fuzzy sets,spherical fuzzy sets with linguistic 2-tuples are constructed respectively,and the rationality is studied.At the same time,the reverse conversion process of normalized linguistic 2-tuples with multi-type data is analyzed.A multi-type fuzzy data normalized algorithm is proposed,and the rationality of the algorithm are illustrated via a reference book recommendation example.2.In order to process the linguistic-valued evaluation information under different granularity,the conversion method between 2-tuple linguistic formal context of different granularity is studied.Firstly,combining the 2-tuple linguistic model with the formal concept analysis,2-tuple linguistic formal context and 2-tuple linguistic concept lattice are proposed.At the same time,for the formal contexts with different type of data,a normalized method of multi-type fuzzy formal context is proposed based on the multi-type fuzzy data conversion model.For the normalized formal context,the conversion method between 2-tuple linguistic formal context of different granularity is studied by keeping the directed normalized distance from linguistic terms to central language terms unchanged.3.In order to extract rules with linguistic-valued information,a rule extraction method based on 2-tuple linguistic formal decision context is proposed.Firstly,a quadratic completing method based on 2-tuple linguistic object similarity is studied for the incomplete2-tuple linguistic form context with information loss.Secondly,based on conditional attributes and decision attributes,the 2-tuple linguistic formal decision context is proposed,and rules are extracted according to the fineness relationship between corresponding concept lattices.On this basis,a recommendation algorithm based on 2-tuple linguistic formal decision context is given,and a product recommendation example is used to illustrate its validity and practicality.