Dear Himani,

Fuzzy membership is the process of assigning the fraction values given to each element to identify its membership in different classes such that the sum of all these fraction values is equal to 1. So, if you have N classes and M elements then each m belongs to M has fraction values from F1 to Fn such that F1+ F2+...+ Fn = 1.

In this case you have to select the maximum value from F1 to Fn to identify the class membership for each m in M. After that you can have a vector of size MX1 instead of MXN. This vector can be supplied to SVM, RF ANN or any other algorithm.

To understand fuzzy you need a control background.

Please read my book

The American University Laboratories For Electrical Engineering Part 1

the control section is a great help.

Dear Himani Deshpande,

For this, you need to first understand what is conventional time series and fuzzy time series. In conventional time series, the elements are represented as normal values but in fuzzy time series, the elements are fuzzy elements.

Then you need to understand what is membership value ? The membership degree used to represent each class of an element is called membership value. Always the membership degree should be with in the range [0,1].

Please go through the recently my recently published article " High-Order Fuzzy Time Series Forecasting by Using Membership Values Along with Data and Support Vector Machine " available in RG. Please download it and go through once, so that your doubt will clear.

Further, if you have any doubt feel free to discuss this.

Thanks