Currently, personal data is being collected as big data, and it is believed that many problems can be solved by utilizing the analysis results. Privacy protection is important in the collection and utilization of such data. One of the techniques to utilize data while protecting personal privacy is local differential privacy. In local differential privacy, the data provider itself adds noise to the data and gives it to the provider, so that the provider can use all the collected noisy data for histograms, etc., while protecting the privacy of the data. In the existing local differential privacy, data is treated uniformly regardless of its attributes. Therefore, the usage of noisy data is limited. In this paper, we extend the concept of local differential privacy and propose a new concept of local differential privacy that can be used for various purposes.