Improved bass model using sales proportional average for one condition of mono peak curves

Abstract: "This study provides a modified Bass model to deal with trend curves for basic issues of relevance to individuals from all over the world, for which we collected 16 data sets from 2004 to 2022 and that are available on Google servers as "google trends". It was discovered that the Bass model did not forecast well for curves that have a mono peak with a sharp decrease to some level then have semi-stable with small decrement sales for a long time, thus a new parameter based on r1 and r2 (ratios of average sales) was introduced, which improved the model's prediction ability and provided better results. The model was also applied to a data set taken from the Kaggle website about a subscriber digital product offering for financial services that include newsletters, webinars, and investment recommendations. The data contain 508932 data points about the products sold during 2016-2022. Compared to the traditional Bass model, the modified model showed better results in dealing with this condition, as the expected curve shape was closer to real sales, and the sum of squares error (SSE) value was reduced to a ratio ranging between (36.35-79.3%). Therefore, the improved model can be relied upon in these conditions."