RavenPack

When it comes to using large language models to gain an investment edge, choosing the right data is critical. We've been testing a novel approach that uses open-source language models like FinBERT fine-tuned with RavenPack Annotations a dataset of training sentences annotated with sentiment information on the entities detected in each sentence. This method has shown promising results, 𝐚𝐜𝐡𝐢𝐞𝐯𝐢𝐧𝐠 𝐚𝐧 𝐚𝐯𝐞𝐫𝐚𝐠𝐞 𝐈𝐧𝐟𝐨𝐫𝐦𝐚𝐭𝐢𝐨𝐧 𝐑𝐚𝐭𝐢𝐨 𝐢𝐧𝐜𝐫𝐞𝐚𝐬𝐞 𝐨𝐟 48% 𝐢𝐧 𝐜𝐨𝐧𝐭𝐫𝐨𝐥𝐥𝐞𝐝 𝐭𝐞𝐬𝐭𝐬. Download the White Paper to learn more: https://lnkd.in/dxD9nqEi

The financial industry is buzzing about large language models and their potential to predict stock prices. Peter Hafez, Chief Data Scientist at RavenPack, offers a balanced perspective. He acknowledges the potential of LLMs while highlighting the complexities involved. Read his analysis here: https://lnkd.in/dFYi5BB5

For more than 20 years, RavenPack has been a cornerstone for some of the world’s largest institutional investors, with its data and analytical solutions utilized by the most successful hedge funds and banks. This new funding will be used to bring 𝐁𝐢𝐠𝐝𝐚𝐭𝐚 to market, further enhancing its capabilities. Commenting on the funding, Armando Gonzalez, CEO and Co-Founder of RavenPack, explained: "This investment marks a pivotal moment for RavenPack. With the support of GP Bullhound, we are set to bring our most advanced solution to the market, setting a new standard in the financial services industry. Financial professionals want to tap into the tremendous power of AI, but current chatbots have yet to fully meet their evolving needs. 𝐁𝐢𝐠𝐝𝐚𝐭𝐚.𝐜𝐨𝐦 aims to change that by empowering its users with real-time trusted data, providing a gateway to instant knowledge and confident decision-making for an eager market."

Navigating emerging trends and alt data insights can be a challenge. Stay in the know with The RavenPack Brief, only on LinkedIn. In this edition: • How CEO's positive portrayal during M&A announcements influences investor responses. • Do unique questions during earnings calls yield valuable insights for analysts? • How earnings releases cool off sentiment-driven expectations. • A new model uses Job data to predict stock performance. Access the most important articles, stories, and insights to keep you well-informed.

RavenPack's Chief Product Officer, Kevin Crosbie, will be speaking at the "AI, Data, and Cloud - Better Together" event in Madrid on July 4th. Kevin will explore how AI unlocks actionable insights from data, empowering better decision-making. This event is hosted by AWS Experience Iberia (AWS's customer engagement center in Madrid) and organized by Deloitte. Join us for an insightful discussion: https://lnkd.in/gkY7rv9n

Wolfe Research, LLC has released 𝐚 𝐧𝐞𝐰 𝐬𝐭𝐮𝐝𝐲 𝐭𝐡𝐚𝐭 𝐚𝐧𝐚𝐥𝐲𝐳𝐞𝐬 𝐡𝐨𝐰 𝐠𝐥𝐨𝐛𝐚𝐥 𝐣𝐨𝐛 𝐩𝐨𝐬𝐭𝐢𝐧𝐠𝐬 𝐢𝐧 𝐑𝐚𝐯𝐞𝐧𝐏𝐚𝐜𝐤 𝐉𝐨𝐛 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 𝐡𝐞𝐥𝐩 𝐩𝐫𝐞𝐝𝐢𝐜𝐭 𝐜𝐨𝐦𝐩𝐚𝐧𝐲 𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐨𝐧. They identified a correlation between hiring trends related to skills or locations and future stock performance. Check out our breakdown of their methodology and key findings: https://lnkd.in/dRYHk9UF

In a bid to understand the progress towards the UN's Sustainable Development Goals (SDGs), Amundi, one of Europe’s leading asset managers, leveraged RavenPack's powerful news analytics. Their analysis revealed a critical gap in "SDG 17: Partnerships for the Goals" – a significant step for achieving most SDGs. Read Amundi's full analysis to learn more:

This just in: WatersTechnology has named RavenPack Best Alternative Data Provider at the 2024 WatersTechnology Asia Awards! This recognition acknowledges our leadership in turning heaps of textual data into insights that traders use across asset classes worldwide to capture alpha, mitigate risks, and enhance their viewpoints. RavenPack continues to innovate in generative AI for finance so analysts and investors can enhance their workflows, and explore new ways to work. Thank you to all our RavenPackers who always go the extra mile to deliver the industry reference for alternative data at scale: https://lnkd.in/d9NrRdgx

As large language models (LLMs) advance, organizations must find a balance between experimentation and controlled implementation for responsible AI adoption. Peter Hafez, RavenPack's Chief Data Scientist, recommends a hybrid approach. By combining LLMs with smaller language models (SLMs), companies can enhance trust, transparency, and domain-specific expertise. Read his analysis here: https://lnkd.in/dqZdeKtR

Selecting and timing country-level investments can be challenging as most national economies are deeply interconnected. Our research explores 𝐡𝐨𝐰 𝐦𝐚𝐜𝐡𝐢𝐧𝐞 𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐜𝐚𝐧 𝐡𝐞𝐥𝐩 𝐧𝐚𝐯𝐢𝐠𝐚𝐭𝐞 𝐭𝐡𝐢𝐬 𝐜𝐨𝐦𝐩𝐥𝐢𝐜𝐚𝐭𝐞𝐝 𝐥𝐚𝐧𝐝𝐬𝐜𝐚𝐩𝐞 𝐚𝐧𝐝 𝐟𝐢𝐧𝐝 𝐛𝐞𝐭𝐭𝐞𝐫 𝐫𝐢𝐬𝐤-𝐚𝐝𝐣𝐮𝐬𝐭𝐞𝐝 𝐫𝐞𝐭𝐮𝐫𝐧𝐬 𝐢𝐧 𝐆10 𝐞𝐪𝐮𝐢𝐭𝐢𝐞𝐬. The model, trained on macroeconomic news sentiment, optimizes country allocation for mid-frequency trading. Backtesting shows an Information 𝐑𝐚𝐭𝐢𝐨 𝐨𝐟 0.84 𝐭𝐨 0.61 when holding for 2 days to a week, and an increase in next-day equity returns by 324 𝐛𝐚𝐬𝐢𝐬 𝐩𝐨𝐢𝐧𝐭𝐬 for a 2-day holding period. Get the full research here: https://lnkd.in/dHQQ35uh