Price Action Lab Blog

Premium Market Analysis

DLPAL software

Announcing The Release Of DLPAL PRO v2.0

We are pleased to announce the release of v2.0 of DLPAL PRO. This new version offers generation, updating and creation of train/test and scoring files with just a few mouse clicks. The generated files can be used with machine learning classifiers for the development of directional and long/short strategies.

Version 2.0 of DLPAL PRO is the result of massive extension of the capabilities of v1.0. This new version allows generating and updating files with historical values of attributes that can be used with machine learning algorithms for the development of a variety of trading strategies. The new features include:

  • Creating and then updating historical attribute files with new instances as new data becomes available for a universe of securities.
  • Creation of train/test and score files for a universe of securities from the historical attribute files
  • Saving and loading of default settings
  • Attribute and train/test files maintenance

Below is a screenshot of the updated function in DLPAL PRO v2.0:


Files with historical attribute values have the following header

Date, Open, High, Low, Close, PLong, PShort, Pdelta, S

where P-Long and P-Short are the long and short directional bias, P-delta is P-Long – P-Short and S is the significance of the bias.

The files have the same name of the original data file and extension .pih

Train files have the following header

Date, Open, High, Low, Close, PLong, PShort, Pdelta, S, target

The target is a binary class of 0 or 1 based on the sign of the return one period in the future. For example, in the case of EOD data, 1 means that the close of next day is higher than the close of today and the reverse for 0.

Scoring files have the following header

Date, Open, High, Low, Close, PLong, PShort, Pdelta, S

Scoring files have only one instance and there is no entry for the target. The objective of machine learning is to calculate the probability of the outcome being of class 1 and fill in the value.

Train/test files have extension .pit and scoring files .pis

Fully automated function

All new functions in v2.0 are automated. The user has to supply the historical data for the securities (Date, O, H, L, C) and the program will generate the initial historical files with the attribute values for the number of instances specified (history length.) Then, after the historical data are updated – usually on a daily basis after the market close – all the user has to do is to click Update history files and then Create train and score. The generated files can be used with any language or platform that facilitates machine learning to get the scores and rank securities accordingly.

The edge of DLPAL PRO v2.0

The quality and accuracy of results from machine learning depends on the attributes used. The difficult part is identifying attributes that have economic value. Applying machine leaning classification is the easy part. Although most work has been focused on machine learning algos, very little has been presented on how to generate high quality attributes (feature engineering) with economic value. In fact, the edge is in the attributes. DLPAL PRO calculates values for attributes based on the p-indicator. The values of this indicator result from sophisticated unsupervised/supervised learning. The attributes that are based on the p-indicator offer an edge in detecting short-term price action anomalies. Those who exploit the edge fast may gain economic benefit over those who rely on strategies that have lost their edge.

In the next article, we will include the flow and a specific example for a long/short strategy based on liquid ETFs.

If you have any questions or comments, happy to connect on Twitter: @priceactionlab

You can download a demo of DLPAL from this link. There is a demo for DLPAL PRO you can download here. For more articles about DLPAL and DLPAL PRO click here.

Subscribe via RSS or Email, or follow us on Twitter.


Copyright Notice

Comments are Closed

Theme by Anders Norén