OCLC’s Post

The community-informed Reimagine Descriptive Workflows project and resulting report was an opportunity for OCLC to reflect on our own role in repairing systems and workflows. During the development of OCLC’s linked data ecosystem, the WorldCat ontology, our teams conducted an internal diversity, equity, and inclusion review of the ontology prior to release. #OCLCresearch's Merrilee Proffitt reviews how our teams across the organization came together with library professionals to "dream and enact change." "As a result of this process, OCLC is making improvements to the WorldCat ontology," Merrilee writes. "This effort, working in community to ensure that the WorldCat ontology shifts away from systems and structures that were developed during the nineteenth century, demonstrates that this work is complex, nuanced—and worth undertaking." ⤵️ Click below to read more in this #HangingTogether blog. --- #OCLC #PublicLibrary #PublicLibraries #AcademicLibrary #AcademicLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove

Are we at a “Toy Story” moment with linked data? The first feature-length film created entirely with CGI was a game changer, setting a new standard for animated movies. In this #OCLCnext blog, Mary Sauer-Games, Vice President of Global Product Management at OCLC, provides a synopsis of our commitment to linked data — from the initial research groundwork to the long-term strategy — and shares why this could be a similar "Toy Story" moment for library metadata. "What’s important to remember is that moving to linked data is more than a technological upgrade," Mary writes. "It’s a fundamental shift towards a more discoverable, collaborative and dynamic knowledge ecosystem, where libraries are valued contributors. By participating in this change, libraries not only enhance their own operations but also contribute to the broader goal of creating a more informed and connected world." ⤵️ Click below to read more. --- #OCLC #LinkedData #Metadata #PublicLibrary #PublicLibraries #AcademicLibrary #AcademicLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove

Hi friends who got inspired by our linked data leadership blog. Here is the latest ➡️➡️ "Transitioning to linked data does not require a one-size-fits-all solution. (...) Providing flexibility with how libraries experiment with and learn about linked data without the pressure of complete system overhauls" Many thanks Mary!#oclc #linkeddata

Join us on 6 June (9:00 a.m. EDT/3:00 p.m. CEST) for the OCLC-LIBER Europe "Building for the future" closing plenary session. Libraries are navigating a complex landscape of emerging technologies and ever-growing quantities of data, impacting library operations, and transforming service offerings. Responding to these changes is critical, and between November 2023 and April 2024, OCLC and LIBER jointly convened three interactive discussion sessions to closely examine the opportunities and responsibilities libraries face in providing state-of-the-art services in this challenging environment. The 75-minute closing plenary session will offer key insights gleaned from these sessions, synthesized from the numerous small group discussions of transcontinental participants from the LIBER and OCLC Research Library Partnership networks. A moderated panel of LIBER representatives will respond to this overview and reflect upon how these discussions can help research libraries collaboratively plan in these challenging times. ⤵️ Click below to register for this session and learn more about the series. --- #OCLC #OCLCresearch #AcademicLibrary #AcademicLibraries #ResearchLibrary #ResearchLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove

OCLC's latest position paper highlights a linked data strategy for improved library catalogs today and more connected information in the long term. If you're new to the subject and why it's important, this is a good place to start. The brief, 12-page paper articulates the value that linked data can bring to library catalogs, and the tools already available to help ease libraries into creating a more interconnected experience for information seekers. It follows OCLC's path from decades of linked data research, to experiments and prototypes, the creation of more than 150 million WorldCat Entities, and, most recently, to the web application and APIs now available through OCLC Meridian that libraries can use to get started. ⤵️ Click below to read more. #OCLCnews --- #OCLC #OCLCmeridian #OCLCworldcat #LinkedData #Metadata #PublicLibrary #PublicLibraries #AcademicLibrary #AcademicLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove #LibrariesOfInstagram #LibrariansOfInstagram #ILoveLibraries #LibrariesTransform

Dive into the latest findings ⬇️⬇️ #librarin_eu ✔️ #Livinglab may not be a common term, yet libraries shine as hubs of innovation, hinting at a deeper theme. ✔️ Literature delves into #libraries' purpose and "public value," marking their evolving identity. ✔️ A conceptual revolution in libraries: from offering 'space-place' to 'democratic engagement' services. ✔️ This evolution propels libraries into vibrant, collaborative spaces, transforming #librarians’ roles and skills.

𝗡𝗘𝗪 𝗼𝗻 𝗼𝘂𝗿 𝗯𝗹𝗼𝗴: Delve into the transformative role of #publiclibraries as #livinglabs! Check how the #librarin_eu team explores the experimental realms #libraries are creating for #cocreation and #coinnovation!

I had to pinch myself. I discovered Zeenea half way through writing "The Enterprise Data Catalog" There it was, the technology I thought no one had built. The ideas in my book, materialized. Yesterday, Luc Legardeur published a cartoon showing the ideal behind Zeenea. A data catalog is a search engine, and it should look like one - and be one - it should not be a complicated tool with endless features and poor search power. The illustration is from my book, check out Luc's thoughts+illustration in the comment. #datacatalog #searchengine

One of the most important lessons I’ve learned from Stephen Hardy is to begin any problem with asking good questions - and then listening. As CEO of mySidewalk, he has continued to push all of us to listen to our customers and the experience they need. That’s why I’m so excited for today. We’ve been listening to people like Lourdes Aceves, Cyndi Conn and Isaac Castillo - to make data analysis dead simple - and to help them tell compelling stories about the places they care about. Today, we’re unveiling Sidekick, an AI analyst that helps our customers: - Make a map in a minute. #AskSidekick - and right before your eyes, watch it happen. - Surface key datasets among 5B indicators. #AskSidekick - and the data we’ve curated with our team of data scientists are now at your fingertips. (PS - we’ll also tell you where the data is from and context for what you’re looking at) We know we still have work to do. But we’re listening. And with the partners we have, we know this is just the beginning. https://lnkd.in/gZGyGJgj

🌲 Mastering Decision Trees: The Art of Pruning 🎓 In the intricate world of decision trees, the recursive approach often leads to complex structures, memorizing the training data and falling prey to the infamous overfitting. Enter Chapter 4: Pruning Decision Trees, the strategic antidote to this conundrum. 🚧 The Dilemma: The recursive approach subdivides training cases until every subnode consists of a single class or there are no more attributes to build the tree. Result? A complex structure prone to overfitting. 🌱 Pruning Defined: Pruning is not merely chopping the tree down in favor of a leaf. It's an art, the removal of tree parts that don't contribute to classifying unseen cases. 🔪 Two Pruning Strategies: 1️⃣ Stopping or Prepruning: A decision to halt further division based on a threshold value. Beware of pitfalls—it may stop prematurely or lead to minimal simplification. 2️⃣ Postpruning: A more deliberate, slower approach. Grow the entire decision tree and then meticulously prune the overfitted subtrees. Our aim? Reduce predicting error rates. 🎯 Reducing Errors:   - Predicting Error Rates: Estimate errors from the tree and its subtrees using an unseen set of cases. Techniques like Cost Complexity Pruning and Reduced Error Pruning fall under this family.   - Pessimistic Pruning (C4.5 Approach): Calculate the resubstitution error rate (E/N) for a leaf. This is essentially finding the upper limit of the confidence limits for the binomial distribution, equating it to the predicted error rate. The decision to prune hinges on comparing errors before and after pruning. 🧠 The Wisdom in Pruning: It's not just about simplifying; it's about crafting a tree that stands strong against unseen challenges. The delicate balance between complexity and accuracy. 🌲 Code on GitHub: https://lnkd.in/gBm9TrBS Excited to explore the nuances of pruning decision trees! Let's keep refining our understanding of ML & DS together. 🚀💻 #DecisionTrees #MachineLearning #PruningTechniques #DataScience