Hello,

I am sharing an interesting problem statement to predict the future of stock price based on a multivariate analysis.

As most of the stocks are undervalued. The usual price and volume indicators are not enough to predict the future of any stock performance in 3-6 months time frame.

Therefore, we want to combine the finical reports indicating company performance, of the public equity firms, institutional buy and sale trends( number of stocks bought and sold by institutions),along with the daily price trends of the stock and the index.

My Hypothesis is that the future performance of any stock is strongly impacted by the overall performance stocks index, the trend of financial performances/ KPIs and, institutional holding patterns.

In order to present a price / return trends, I need to find a right model using aforementioned variables along with the daily price trends and establish the model offers better prediction over standard ARIMA / time series models.

However, the challenge is that the frequency of such data points are not uniform. The financial records are released every quarter while the institutional holdings are released on a monthly basis. I can get the daily price trends of stocks and index without any problem.

Therefore, please advice me on how to prepare the data model to fil the gaps with the non availability of daily data for the other variables?

Current data points are as follows :

Please let me know how do address the data preparation issue and identify a set of models that can help us achieving a satisfactory result.

Thanks and Regards,

Prasenjit Bhadra