High Frequency trading dataset copyright FirstRateData.com

What's new:

Add tick dataset :)
Add transaction fee
The model needs to learn how to avoid the cost from transaction fee, which means it should avoid buying too many times
You can add a supplimentary model for Qnet (No consideration for transaction fee), and let it consider the transaction cost
A trail model will be: Use a LSTM and input action and output the same way with loss = loss-transaction fee
The model simply decide whether to execute this order or just stay. Buy and sell are determined by Qnet
Add drop trend dataset

Abstract copyright UK Data Service and data collection copyright owner.

The aim of this study is to contribute to decision theory and to provide information valuable to management in finance. This research seeks to clarify and measure the decision styles of traders, whose work demands quick and balanced judgement under conditions of risk and uncertainty. A valid taxonomy of individuals' psychological preferences and decision style and a systematic analysis of which behaviours are likely to occur under what conditions for which individual could aid selection, placement and management systems.
The aim of the project was to develop a new measure based on previous research which indicated that decision style is likely to compromise a number of psychological constructs. The measure would focus on risk dispositions, cognitive orientations and emotional involvement. Two further measures would collect data concerning the organisational-level processes of recruitment and placement of new employees and measure individual-level performance. Longitudinal performance data would also be gathered to examine change over time.
The sample comprised investment banks and fund management companies. Data would be gathered from firms with different roles in the industry to enable similarities and differences across firms to be examined. Feedback from the project would have academic and practical value, taking the form of academic papers and company-specific reports.
The dataset held at the Archive does not contain all the data collected during the project - some restricted-use personality test data are not included.
Main Topics:

The dataset contains data relating to 118 traders sampled from four investment banks. The data represent a number of issues which influence the decision-making process and trader performance. Six different methods of data collection were used:
biographical information;
personality data gathered using a standardised measure (these data are not included in the dataset held at the Archive);
a newly-developed measure (Risk Assessment Tool);
self-ratings on four aspects of performance;
managers' ratings of traders on the same four aspects of performance and questionnaire data relating to perceptions of pay.

The data enable assessment of relationships between a number of individual difference factors.

For further details about the variables, please see documentation.