Fuzzy Clustering-Based Ensemble Approach to Predicting Indian Monsoon

Abstract EN

Indian monsoon is an important climatic phenomenon and a global climaticmarker.

Both statistical and numerical prediction schemes for Indian monsoonhave been widely studied in literature.

Statistical schemes are mainly basedon regression or neural networks.

However, the variability of monsoon is significant over the years and a single model is often inadequate.

Meteorologists revisetheir models on different years based on prevailing global climatic incidents likeEl-Niño.

These indices often have degree of severity associated with them.

In thispaper, we cluster the monsoon years based on their fuzzy degree of associativityto these climatic event patterns.

Next, we develop individual prediction modelsfor the year clusters.

A weighted ensemble of these individual models is usedto obtain the final forecast.

The proposed method performs competitively withexisting forecast models.