Estimating Probability Estimate of low sample data

by Baktaawar   Last Updated September 13, 2018 10:19 AM

I have a problem pertaining practical real-life situation we encounter quite often.

Lets say I have a e-commerce site selling clothes. There is a particular shirt which was sent N units to the warehouse and let's say few of them were bought online.

SO here are the questions:

1). What would be the estimate of future probability of selling of this shirt and the no. of units we can estimate to be sold in future.

2). If the no. of units are lets say 10 and we sold 10 of them, this is a low sample problem and sample estimate here is like 1.0, but that can't be correctly used to estimate the probability parameter. How to estimate the future probability of selling of this shirt.

3). How to get an estimate for a completely new product which doesn't have any sales yet.

4). In general if someone buys 5 products. Normally if the items are independent of each other we can estimate the sales by taking the sum of probabilities times the no. of units etc. But in practice items are not bought independently, hence the probability distribution of each of these 5 items is not independent. In this case how can one get an estimate



Related Questions



Making predictions with Bayesian models

Updated February 22, 2017 16:19 PM

How to Apply Bayes Theorem in a Poll?

Updated September 03, 2018 15:19 PM

Which model should I choose for my ARIMA time series?

Updated October 10, 2017 22:19 PM