The purpose of this is to calculate the moving average with Holt's linear method, a double exponential smoothing model where moving averages depend on window size and the alpha, beta∈ (0,1]. alpha and beta parameters default to 0.3 and 0.1, respectively. The smaller alpha value will produce more smoothing and more lag. A smaller beta value will emphasize long-term trends. The rest of the aggregation for the holt model is as follows:

"weekly_moving_avg": {"moving_avg":{"buckets_path":"weekly_change","model":"holt","settings":{"alpha":0.3, "beta":0.3},"window":4,"predict":4}}}}