Dynamic Fuzzy Machine Learning

DFSSL seeks to obtain a function $\left(\overleftarrow{f},\overrightarrow{f}\right):\left(\overleftarrow{X},\overrightarrow{X}\right)\to \left(\overleftarrow{Y},\overrightarrow{Y}\right)$ that can accurately predict the label $\left(\overleftarrow{y},\overrightarrow{y}\right)$ of sample $\left(\overleftarrow{x},\overrightarrow{x}\right)$ and classify it into class $\left({\overleftarrow{C}}_y,{\overrightarrow{C}}_y\right)$ with membership degree $\left({\overleftarrow{C}}_y\left(\overleftarrow{x}\right),{\overrightarrow{C}}_y\left(\overrightarrow{x}\right)\right).$ Here, $\left({\overleftarrow{x}}_i,{\overrightarrow{x}}_i\right)\mathrm{and}\left({\overleftarrow{x}}_i,\overrightarrow{x}{\text{'}}_i\right)$ are d-dimensional vectors, $\left({\overleftarrow{y}}_i,{\overrightarrow{y}}_i\right)\in \left(\overleftarrow{Y},\overrightarrow{Y}\right)$ is the label of sample $\left({\overleftarrow{x}}_i,{\overrightarrow{x}}_i\right),\mathrm{and}\left|\left(\overleftarrow{L},\overrightarrow{L}\right)\right|,\left|\overleftarrow{U},\overrightarrow{U}\right|$ are the sizes of $\left(\overleftarrow{L},\overrightarrow{L}\right)\mathrm{and}\left(\overleftarrow{U},\overrightarrow{U}\right),$ respectively (that is, the number of samples contained in the labelled and unlabelled sample sets).

Compared with the previous definitions, ...