Stock Market Prediction using Darvas Box Technique

The reason behind this paper is to showcase an attempt to create an algorithm that will implement the Darvas Box strategy and automatically identify profitable stocks. In addition, it will also recommend a stop loss order to the user so that he/she procures minimum loss. Our datasets consist of 3 month daily company-wise historical data for companies featured on the NSE 52 week high list. Our datasets have been gathered with a particular focus on the following features: Last Traded Price (LTP) and the volume of the stocks traded. The algorithm is particularly targeted towards short to medium term delivery trading. The algorithm works best when the analysis is done once the market closes. This is purely based on technical analysis of the stocks. To counter the unforeseen changes due to uncontrollable circumstances there is a stop loss applied to each stock that follows the pattern so as to minimize the losses occurred during the trades.