Due to the pandemic's intensification, it was estimated that online sales will account for 18.1% of all retail sales globally in 2021. The days of physically checking items in stores, exploring multiple locations, and searching for the appropriate item in the right place are long gone. According to statistics, many store customers use cell phone devices to check prices. As a result, online retailers need to set the proper pricing for their products. Knowing how much a product is worth can take time and effort. Minute details greatly influence the price. Considering the volume of things being sold online, pricing products is increasingly challenging at scale. Seasonal changes in clothing pricing trends differ from those in electronic product prices, which depend on the technology and features of the product. Additionally, brand recognition significantly impacts product pricing [1]. The field of recommendation systems is rapidly developing in web services and e-commerce applications. Searching through various products takes a lot of time when purchasing online. A recommendation system helps expedite the discovery of an extensive range of goods that clients are interested in. As it is simple and dependable for a consumer to purchase online and locate the ideal solutions for them without any hassles, the use of this effective suggestion system is growing day by day. [2] We are utilizing the Mercari Dataset, Japan's largest community-powered shopping app. We would develop an algorithm that automatically recommends the appropriate product prices. The vendor and the buyer will benefit from this in terms of pricing, which should result in more transactions—product descriptions provided by customers that include information on the product category, brand, and condition. The suggested model includes dynamic pricing and recommended prices for the products. The following machine learning algorithms are used in this model: Decision Tree Regressor for tree-based models, Ridge Regression to reduce training error while balancing model complexity, Lasso Regression to reduce squared loss with l1 regularisation, and Light GBM, which has a quick learning time, high efficiency, good accuracy, and good compatibility with large data sets. To improve the efficiency Deep Learning Algorithm, simple RNN is used to compare. (Abstract)

E-commerce, Dynamic pricing, Machine learning algorithms, Lasso regression, Ridge Regression, Decision Tree Regressor, Light GBM, Deep learning algorithm, RNN