Tag: predictive models

The science of turning historical data into actionable insights is far from magic. And while organizations have successfully used predictive analytics for years, we're in the midst of a transformation. New tools, vast amounts of data, enhanced computing power and decreasing implementation costs are making predictive analytics increasingly accessible. And business leaders from varying industries and functions can now use the outcomes to make strategic decisions and manage risk. What is predictive analytics? Predictive analytics is a type of data analytics that uses statistical modeling and machine learning techniques to make predictions based on historical data. Organizations can use predictive analytics to predict risks, needs and outcomes. You might use predictive analytics to make an immediate decision. For example, whether or not to approve a new credit application based on a credit score — the output from a predictive credit risk model. But organizations can also use predictive analytics to make long-term decisions, such as how much inventory to order or staff to hire based on expected demand. How can predictive business analytics help a business succeed? Businesses can use predictive analytics in different parts of their organizations to answer common and critical questions. These include forecasting market trends, inventory and staffing needs, sales and risk. With a wide range of potential applications, it’s no surprise that organizations across industries and functions are using predictive analytics to inform their decisions. Here are a few examples of how predictive analytics can be helpful: Financial services: Financial institutions can use predictive analytics to assess credit risk, detect fraudulent applicants or transactions, cross-sell customers and limit losses during recovery. Healthcare: Using data from health records and medical devices, predictive models can predict patient outcomes or identify patients who need critical care. Manufacturing: An organization can use models to predict when machines need to be turned off or repaired to improve their longevity and avoid accidents. Retail: Brick-and-mortar retailers might use predictive analytics when deciding where to expand, what to cross-sell loyalty program members and how to improve pricing. Hospitality: A large hospitality group might predict future reservations to help determine how much staff they need to hire or schedule. Advanced techniques in predictive modeling for financial services Emerging technologies, particularly AI and machine learning (ML), are revolutionizing predictive modeling in the financial sector by providing more accurate, faster and more nuanced insights. Taking a closer look at financial services, consider how an organization might use predictive credit analytics and credit risk scores across the customer lifecycle. Marketing: Segment consumers to run targeted marketing campaigns and send prescreened credit offers to the people who are most likely to respond. AI models can analyze customer data to offer personalized offers and product recommendations. Underwriting: AI technologies enable real-time data analysis, which is critical for underwriting. The outputs from credit risk models can help you to quickly approve, deny or send applications for manual review. Explainable machine learning models may be able to expand automation and outperform predictive models built with older techniques by 10 to 15 percent.1 Fraud detection models can also raise red flags based on suspicious information or behaviors. Account management: Manage portfolios and improve customer retention, experience and lifetime value. The outputs can help you determine when you should adjust credit lines and interest rates or extend offers to existing customers. AI can automate complex decision-making processes by learning from historical data, reducing the need for human intervention and minimizing human error. Collections: Optimize and automate collections based on models' predictions about consumers' propensity to pay and expected recovery amounts. ML models, which are capable of processing vast amounts of unstructured data, can uncover complex patterns that traditional models might miss. Although some businesses can use unsupervised or “black box" models, regulations may limit how financial institutions can use predictive analytics to make lending decisions. Fortunately, there are ways to use advanced analytics, including AI and ML, to improve performance with fully compliant and explainable credit risk models and scores. WHITE PAPER: Getting AI-driven decisioning right in financial services Developing predictive analytics models Going from historical data to actionable analytics insights can be a long journey. And if you're making major decisions based on a model's predictions, you need to be confident that there aren’t any missteps along the way. Internal and external data scientists can oversee the process of developing, testing and implementing predictive analytics models: Define your goal: Determine the predictions you want to make or problems you want to solve given the constraints you must act within. Collect data: Identify internal and external data sources that house information that could be potentially relevant to your goal. Prepare the data: Clean the data to prepare it for analysis by removing errors or outliers and determining if more data will be helpful. Develop and validate models: Create predictive models based on your data, desired outcomes and regulatory requirements. Deciding which tools and techniques to use during model development is part of the art that goes into the science of predictive analytics. You can then validate models to confirm that they accurately predict outcomes. Deploy the models: Once a model is validated, deploy it into a live environment to start making predictions. Depending on your IT environment, business leaders may be able to easily access the outputs using a dashboard, app or website. Monitor results: Test and monitor the model to ensure it's continually meeting performance expectations. You may need to regularly retrain or redevelop models using training data that better reflects current conditions. Depending on your goals and resources, you may want to start with off-the-shelf predictive models that can offer immediate insights. But if your resources and experience allow, custom models may offer more insights. CASE STUDY: Experian worked with one of the largest retail credit card issuers to develop a custom acquisition model. The client's goal was to quickly replace their outdated custom model while complying with their model governance requirements. By using proprietary attribute sets and a patented advanced model development process, Experian built a model that offered 10 percent performance improvements across segments. Predictive modeling techniques Data scientists can use different modeling techniques when building predictive models, including: Regression analysis: A traditional approach that identifies the most important relationships between two or more variables. Decision trees: Tree-like diagrams  show potential choices and their outcomes. Gradient-boosted trees: Builds on the output from individual decision trees to train more predictive trees by identifying and correcting errors. Random forest: Uses multiple decision trees that are built in parallel on slightly different subsets of the training data. Each tree will give an output, and the forest can analyze all of these outputs to determine the most likely result. Neural networks: Designed to mimic how the brain works to find underlying relationships between data points through repeated tests and pattern recognition. Support vector machines: A type of machine learning algorithm that can classify data into different groups and make predictions based on shared characteristics. Experienced data scientists may know which techniques will work well for specific business needs. However, developing and comparing several models using different techniques can help determine the best fit. Implementation challenges and solutions in predictive analytics Integrating predictive analytics into existing systems presents several challenges that range from technical hurdles to external scrutiny. Here are some common obstacles and practical solutions: Data integration and quality: Existing systems often comprise disparate data sources, including legacy systems that do not easily interact. Extracting high-quality data from these varied sources is a challenge due to inconsistent data formats and quality. Implementing robust data management practices, such as data warehousing and data governance frameworks, ensure data quality and consistency. The use  of APIs can facilitate seamless data integration. Scalability: Predictive business analytics models that perform well in a controlled test environment may not scale effectively across the entire organization. They can suffer from performance issues when deployed on a larger scale due to increased data volumes and transaction rates. Invest in scalable infrastructure, such as cloud-based platforms that can dynamically adjust resources based on demand. Regulatory compliance: Financial institutions are heavily regulated, and any analytics tool must comply with existing laws — such as the Fair Credit Reporting Act in the U.S. — which govern data privacy and model transparency. Including explainable AI capabilities helps to ensure transparency and compliance in your predictive models. Compliance protocols should be regularly reviewed to align with both internal audits and external regulations. Expertise: Predictive analytics requires specialized knowledge in data science, machine learning and analytics. Develop in-house expertise through training and development programs or consider partnerships with analytics firms to bridge the gap. By addressing these challenges with thoughtful strategies, organizations can effectively integrate predictive analytics into their systems to enhance decision-making and gain a competitive advantage. From prediction to prescription While prediction analytics focuses on predicting what may happen, prescription analytics focuses on what you should do next. When combined, you can use the results to optimize decisions throughout your organization. But it all starts with good data and prediction models. Learn more about Experian's predictive modeling solutions. 1Experian (2020). Machine Learning Decisions in Milliseconds *This article includes content created by an AI language model and is intended to provide general information.

Despite the constant narrative around “unprecedented times” and the “new normal,” if the current market volatility tells us anything, it’s to go back to basics. As financial institutions navigate COVID-19’s economic impact, and challenges that are likely to be different or more extreme than in the past, the best credit portfolio management practices are fundamental. The global pandemic impacts today’s data as existing data and analytics may not accurately reflect what is happening now, resulting in inaccurate portfolio assessment. In order to successfully navigate loss forecasting, predicting borrower behavior and controlling loss ratios, lenders must engage new data, analytics and economic scenarios suited for today’s changing times. In Experian’s latest white paper, “Credit Portfolio Management After the COVID-19 Recession,” we’ll explore best practices to combat the following challenges: Forecasting credit losses despite increased economic volatility Businesses have long used a variety of data, analytics and models to anticipate and project the future direction of their organization based on a number of data points; however, with the onset of the global pandemic, long-standing scenarios became suddenly irrelevant.   Predicting borrower behavior given increased financial disparities The post-pandemic and pre-pandemic worlds are very different places for some borrowers. Pandemic-related job losses and other economic effects will not be spread evenly and this variability may be reflected in lenders’ portfolios.   Controlling loss ratios In the post-COVID world, it will be mission critical for lenders to use high-quality and up-to-date data to balance priorities and identify which areas of their portfolio need attention now.   Whether your portfolio is doing better than expected, as expected, or worse than expected, now is the time to refresh portfolio management strategy. Lenders should be watching for early indicators in loan portfolios to better navigate a fluctuating economy and that requires new resources and better tools. Take control of your business’ trajectory. Download now

This is the third in a series of blog posts highlighting optimization, artificial intelligence, predictive analytics, and decisioning for lending operations in times of extreme uncertainty. The first post dealt with optimization under uncertainty and the second with predicting consumer payment behavior. In this post I will discuss how well credit scores will work for consumer lenders during and after the COVID-19 crisis and offer some recommendations for what lenders can be doing to measure and manage that model risk in a time like this. Perhaps no analytics innovation has created opportunity for more individuals than the credit score has. The first commercially available credit score was developed by MDS (now part of Experian) in 1987. Soon afterwards FICO® popularized the use of scores that evaluate the risk that a consumer would default on a loan. Prior to that, lending decisions were made by loan officers largely on the basis on their personal familiarity with credit applicants. Using data and analytics to assess risk not only created economic opportunity for millions of borrowers, but it also greatly improved the financial soundness of lending institutions worldwide. Predictive models such as credit scores have become the most critical tools for consumer lending businesses. They determine, among other things, who gets a loan and at what price and how an account such as a credit line is managed through its life cycle. Predictive models are in many cases critical for calculating loan and loss reserves, for stress testing, and for complying with accounting standards. Nearly all lenders rely on generic scores such as the FICO® score and VantageScore® credit score. Most larger companies also have a portfolio of custom scorecards that better predict particular aspects of payment behavior for the customers of interest. So how well are these scorecards likely to perform during and after the current pandemic? The models need to predict consumer credit risk even as: Nearly all consumers change their behaviors in response to the health crisis, Millions of people—in America and internationally—find their income suddenly reduced, and Consumers receive large numbers of accommodations from creditors, who have in turn temporarily changed some of their credit reporting practices in response to guidelines in the federal CARES Act. In an earlier post, I pointed out that there is good reason to believe that credit scores will tend to continue to rank order consumers from most likely to least likely to repay their debts even as we move from the longest economic expansion in history to a period of unforeseen and unexpected challenges. But the interpretation of the score (for example, the log odds or the bad rate) may need to be adjusted. Furthermore, that assumes that the model was working well on a lender’s population before this crisis started. If it has been a long time since a scorecard was validated, that assumption needs to be questioned. Because experts are considering several different scenarios regarding both the immediate and long-term economic impacts of COVID-19, it’s important to have a plan for ongoing monitoring as long as necessary. Some lenders have strong Model Risk Management (MRM) teams complying with requirements from the Federal Reserve, Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC). Those resources are now stretched thin. Other institutions, with fewer resources for MRM, are now discovering gaps in their model inventories as they implement operational changes. In either case, now’s the time to reassess how well scorecards are working. Good model validation practices are especially critical now if lenders are to continue to make the sound data-driven decisions that promote fairness for consumers and financial soundness for the institution. If you’re a credit risk manager responsible for the generic or custom models driving your lending, servicing, or capital allocation policies, there are several things you can do--starting now--to be sure that your organization can continue to make fair and sound lending decisions throughout this volatile period: Assess your model inventory. Do you have good documentation showing when each of the models in your organization was built? When was it last validated? Assign a level of criticality to each model in use. Starting with your most critical models, perform a baseline validation to determine how the model was performing prior to the global health crisis. It may be prudent to conduct not only your routine validation (verifying that the model was continuing to perform at the beginning of the period) but also a baseline validation with a shortened performance window (such as 6-12 months). That baseline validation will be useful if the downturn becomes a protracted one—in which case your scorecard models should be validated more frequently than usual. A shorter outcome window will allow a timelier assessment of the relationship between the score and the bad rate—which will help you update your lending and servicing policies to prevent losses. Determine if any of your scorecards had deteriorated even before the global pandemic. Consider recalibrating or rebuilding those scorecards. (Use metrics such as the Population Stability Index, the K-S statistic and the Gini Coefficient to help with that decision.) Many lenders chose not to prioritize rebuilding their behavioral scorecards for account management or collections during the longest period of economic growth in memory. Those models may soon be among the most critical models in your organization as you work to maintain the trust of your accountholders while also maintaining your institution’s financial soundness. Once the CARES accommodation period has expired, it will be important to revalidate your models more frequently than in the past—for as long as it takes until consumer behavior normalizes and the economy finds its footing. When you find it appropriate to rebuild a scorecard model, consider whether now is the time to implement ethical and explainable AI. Some of our clients are finding that Machine Learned models are more predictive than traditional scorecards. Early Experian research using data from the last recession indicates this will continue to be true for the foreseeable future. Furthermore, Experian has invested in Research and Development to help these clients deliver FCRA-compliant Adverse Action reasons to their consumers and to make the models explainable and transparent for model risk governance and compliance purposes. The sudden economic volatility that has resulted from this global health crisis has been a shock to all organizations. It is important for lenders to take the pulse of their predictive models now and throughout the downturn. They are especially critical tools for making sound data-driven business decisions until the economy is less volatile. Experian is committed to helping your organization during times of uncertainty. For more resources, visit our Look Ahead 2020 Hub. Learn more