We begin our preliminary modelling assuming that the model, conditioned on the related components, does not depend upon the 12 months of the public sale. The e-public sale statistics (2018-19) consists of the title of tea leaf type, complete lots provided in auctions, whole amount sold in packets and in quantity, and common price. We shall later see in the cross-validation procedure, that our predictions are fairly satisfactory to assert that our assumption was not falsified.
On the other hand, contemplating a classical approach to measure the standard error, we discover that the predicted costs lie between 71.09% and 140.65% of the true prices about 95% of the time. Nonetheless, we couldn’t yet allege that the auctioneers’ valuation had a causal influence on the value stage, though it seemed to be an indispensable predictor. Thus the valuation part is significant for the prediction of the costs, and the method can’t be automated by a simple linear or log-linear mannequin approach. From the easy statistical evaluation with a linear model performed above, it appears Valuation is indeed a pertinent variable in the explanation of the Pricing system.
The histogram exhibits a single modal distribution. We tried fitting a single lognormal distribution, our p-values for the chi-square goodness of match statistic came out as 0.2665 and for One sample Kolmogorov Smirnov check came out as 0.6908, which suggests a reasonably good match. The same sample follows as in Cluster 2, unimodal from histogram, but a single log-regular fits badly (p-worth 2e-04). But match with mixture of two log-normals give fairly properly fits (p-worth 0.3855). Thus we report a mixture of two log-regular distributions with the following properties. Thus we report the ratio for this cluster to observe a an unimodal log-regular distribution with the next properties.
Additionally, from the Q-Q plot, it is clear that there are some doable outliers current at decrease prices. Hence, this pricing mannequin approximately makes an error from 74.43% to 145.86% of the true prices of the tea packets. This is based on a strong strategy to estimate the usual error. Testing the performance of the fitted model on the testing set, the 2.5% quantile of the residuals comes as -0.29539, while the 97.5% quantile of the residuals is 0.37754 in logarithmic scale.