Organizations have access to more data than ever. Making sense of this data and providing actionable insights, however, can be a challenge. As a result, businesses are missing out on the insights needed to drive real business outcomes from data.
Our team of Data Scientists have worked to combine our proprietary XBRL, Artificial Intelligence and Machine Learning technologies to effectively make predictions that link data to business insights. Our latest tool is a preview of this functionality, predicting the likelihood that an SEC filing will require amendment. Our tool looks at historical data and applies statistical analysis to identify patterns. Discovering novel insights, the tool is able to make predictions into the future.
We use a collection of machine learning techniques that examine previous statements and determine the patterns related to subsequent amendments due to misstatement and AAERs.
10-Ks are released annually; 10-Qs are released quarterly.
Companies on the US stock market typically release three 10-Qs per year (Q1, Q2, and Q3) and one 10-K, which includes the Q4 information.
All data used here is publically available from the SEC. Data is collected daily.
We require a set of key financial information to make predictions. If this necessary data is not available or malformed in the XBRL filing, then that organization is omitted from our database. As we collect data on a daily basis, more organizations could be made available once we have the necessary data to make predictions.
Yes. As we gather more data and continue refining the models, our predictions related to any given company may change. As well, given changes in the economy, laws, competition, and so on, the machine learning model predictions will need to change to reflect these.
As of today, the model is trained strictly on public US data. The patterns found may not apply to private companies or companies in other countries and the accuracy will vary. However, it can identify issues - which may not be predictive of misstatement or changes in ratios - but may nevertheless be concerns. As more data becomes available, the models can potentially be applied to private companies around the world.