The Brain Sentiment Indicator (BSI) measures the “mood” on more than 10000 global stocks based on the analysis of financial news using Natural Language Processing techniques.
As a free trial we provide approximately four years of history for a subset of about 3000 stocks and upon request a 30 days access to the daily report corresponding to the full universe.
Brain Machine Learning proprietary platform is exploited to generate a daily stock ranking based on the predicted future returns of a universe of 1000 stocks on four time horizons: 2, 3, 5, 10 and 21 days (other time horizons could be developed and tested upon request). The model implements a series of techniques to reduce the well-known overfitting problem for financial data.
Brain Market Sentiment (BMS) provides a daily a score on the general mood of the market by automatically clustering by topic thousands of news from most popular financial media. The sentiment of each topic is calculated using Brain proprietary Natural Language Processing platform.. The BMS provides an aggregate score for the news topic sentiment of the current day.
Risk ON / Risk OFF signals based on VIX statistical indicators (Dynamic Volatility Signal) and including measures of financial stress indicators and macro-economic environment (Brain Dynamic Allocation Indicator). Portfolio strategies based on these signals are implemented: the strategies toggle between two dynamic portfolios, each of which is monthly rebalanced.
With Unsupervised Machine Learning techniques our algorithms identify non trivial patterns among a large number of financial and macroeconomic data to find days in the past which are “similar" to the current scenario. Investment models can be built by analyzing the performance of various financial instruments on the market clusters identified by the system.
Algorithm-based selection among a large database of companies of a basket of stocks whose business is related to a specific theme (es. "nanotechnology"). The selection is performed by analyzing company public documents and web pages by leveraging on natural language processing and machine learning classifications and clustering techniques.
The Brain Language Metrics (BLM) on Company Filings dataset has the objective of monitoring several language metrics on 10-Ks and 10-Qs company reports for 6000+ US stocks. Some examples of such metrics are financial sentiment, percentage of words belonging to financial domain classified by language types (e.g. “litigious” language), similarity between documents. The analysis is available for the whole report and for specific sections (e.g. Risk Factors and MD&A).
Brain has developed a method to analyze the sentiment and various language complexity metrics of earnings calls transcripts. Each call is divided in the main parts (Management, Q&A) and actors (Managers, Analysts); for each piece the metrics are calculated and a correlation score is evaluated between such metrics (sentiment, readability scores, lexical metrics) and the average future return of a company.
Brain Market Monitor Dashboard allows the monitoring of markets through Brain proprietary signals and a snapshot of Brain alternative datasets.
This a very useful tool for the investor to augment its awareness in the decision process with a complementary view to common market data.
Brain leverages its proprietary Machine Learning infrastructure for the validation of alternative datasets; given a new dataset we are going to integrate this into our existing machine learning model for stock ranking and during the procedure we will evaluate a series of validation metrics to assess if the new data brings alpha.
Brain products and solutions leverage on Natural Language Processing techniques (NLP) to extract from structured and unstructured texts meaningful metrics such as sentiment, language complexity and topics. In the context of NLP we use various machine learning techniques to assess the relevance of a company document (e.g. text extracted from web site) with respect to a specific theme (es. “nanotechnology” or “robotics”) or to identify the relevant topics in documents.
Brain has developed a set of Machine Learning and financial features engineering tools aiming at providing inference on the markets. Our models yield statistical predictions on targets such as assets returns; using ensemble machine learning models we can calculate probabilities associated to the spectrum of predictions. These tools can be used as building bricks for investment strategies or for proprietary and third parties’ portfolio models.
Brain combines various clustering algorithms together with dimension reduction techniques to extract relevant features and to cluster various types of data sets, for example all company documents by topic or the past history of market days in order to extract meaningful information.
Brain has developed a proprietary backtesting and validation approach that we use to test and optimize our models, so that our results are less dependent on the specific historical trajectory markets have undergone. The method can be used also to validate or to optimize third parties’ models.
Brain assists Investment Management firms in the development of their proprietary algorithms.
[29/05/2020 - IEX Cloud Blog] Current times are indeed turbulent again, and for better or worse, financial markets are displaying their high potential to react to these changes on a global scale. Given this new reality, when it comes to data, it's increasingly valuable to be able to identify and process new information, spot emerging relevant topics, and assess their potential impact on financial markets and the economy. To achieve this, investors are ...
[16/12/2019 - Crux Informatics Blog] Brain is a research company that develops proprietary signals and algorithms for investment strategies. Brain also supports clients in developing, optimizing and validating their own proprietary models.
The Brain platform includes Natural Language Processing (NLP) and Machine Learning (ML) infrastructures which enable clients to integrate state-of-the-art approaches into their strategies ...
[SSRN paper by Matúš Padyšák, Quantpedia.com] This research studies the similarity of language used in the filings using data which enables to analyze what type of language is similar. Results show that the similarity of the positive language is the most profitable option. From a practical point of view, the positive similarity effect is examined. Results show that the lowest positive similarity stocks significantly outperform the highest positive similarity stocks. The effect cannot be explained by the common asset pricing models ...
[SSRN paper by D.F. Ahelegbey, P. Cerchiello, R. Scaramozzino, Dept Economics, Pavia University]
How much the largest worldwide companies, belonging to different sectors of the economy, are suffering from the pandemic? Are economic relations among them changing? In this paper, we address such issues by analysing the top 50 S&P companies by means of market and textual data. Our work proposes a network analysis model that combines such two types of information to highlight the connections among companies ...