Researchers Submit Patent Application, “Systems And Methods For Collecting And Processing Alternative Data Sources For Risk Analysis And Insurance”, for Approval (USPTO 20220005125): Patent Application

2022 JAN 26 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- From Washington, D.C., NewsRx journalists report that a patent application by the inventors Andrews, Geoffrey R. (Goleta, CA, US); Drucker, Max Leslie (Santa Barbara, CA, US), filed on September 16, 2021, was made available online on January 6, 2022.

No assignee for this patent application has been made.

News editors obtained the following quote from the background information supplied by the inventors: “Insurance businesses depend critically on their ability to assess risk, such as to predict the likelihood and extent of the loss events that they cover in a given type of policy, so that they can set prices appropriately. Insurance companies use risk models that incorporate information known about policyholders to predict levels of risk, pricing models that set prices appropriately given the levels of risk, and other models to support a range of operations across the lifecycle of a policy or claim. The models have been populated by traditional data sources, such as information obtained during the application process (including information reported by applicants, as well as information derived from sources like credit reports and medical histories), information about loss histories of individual and aggregate policyholders, and demographic and actuarial information about a population or group of similar policyholders. Insurance companies face endemic challenges, such as moral hazard (the tendency of policyholders to engage in riskier behavior once insured) and adverse selection (the tendency of riskier policyholders to choose insurance), such that classifying prospective policyholders by their relative risk characteristics is critical to profitability. Also, insurance companies face challenges in setting prices based on risk. If prices are set too high for the level of risk, policyholders will decline to purchase coverage, reducing revenues and profits. If prices are set too low, risky policyholders will purchase too much coverage, resulting in potentially catastrophic loss levels. Accordingly, insurance companies always need improved methods and systems for predicting and assessing risk. Historically, the information about a policyholder’s characteristics and behavior has been difficult to obtain outside the application process; however, the Internet has led to a proliferation of information about individuals and businesses. This information includes self-published information, such as social media posts, blogs and the like, as well as business websites, publications and advertising materials. The information also includes information posted by others, such as third-party rating sites, social media sites and postings about individuals by third parties, and many others. The volume of such information is extensive, and the reliability is uncertain, but insurance companies cannot afford to ignore these sources at the risk of missing relevant data for assessing and pricing risk. A need exists for methods and systems for collecting, organizing, processing, and analyzing alternative data sources for the benefit of a wide range of operating activities of insurance companies.

“In the case of insurance activities involving individual policyholders, there is a voluminous amount of data publicly available online, with individuals contributing additional data each second. For example, users may provide information via social networking websites, online commerce websites, media sharing websites, news websites, and many other types of mechanisms. However, given the volume of information, the ease of creating “new” web identities, the relative anonymity or lack of verification of data, and other similar reasons, it can be challenging to determine what data may be accurately attributed to a particular individual.

“When attempting to evaluate risk associated with an individual, the voluminous information available online, if analyzed properly, could provide valuable insights. Furthermore, these insights may not be available via other means. For these reasons and others, it is desirable to develop a system and method to apply predictive social scoring to perform focused risk assessment. Aspects of the present disclosure fulfill these and other desires.”

As a supplement to the background information on this patent application, NewsRx correspondents also obtained the inventors’ summary information for this patent application: “A significant amount of information is available to and needed by businesses, such as insurance providers and the like, to provide competitive and timely services. This includes, among many other types, information that is relevant to assessing risk, such as for underwriting insurance policies, setting prices, determining appropriate coverage amounts, and other activities. Information can be sourced from a range of sources, including the Internet and the like that does not always conform to the technical requirements for providing such services. Additionally, the information available from sources, such as websites and the like on the Internet can be conflicting and/or misleading, often across and within such sources. For effective use of such information, methods and systems are presented herein to, among other things, resolve conflicts, clear up ambiguity, divide and categorize, and quantify non-numeric content (e.g., free form text, descriptive content, images, and the like) so that the information can be used effectively (such as by increasing confidence in risk assessments) when performing services and the like based thereon. One exemplary use of such methods and systems involves generation and/or validation of a risk index associated with an offer for service, such as an insurance service, to target clients and the like. Another exemplary use of such methods and systems involves generation and/or validation of other risk indices, such as indices of risks of certain activities (such as specific business activities), services, or the like, such as the risk involved in a business offering a particular service or product, such as serving alcohol. Other exemplary uses and methods and systems are provided herein for collecting, organizing, processing, and analyzing alternative data sources for the benefit of a wide range of operating activities of insurance companies.

“In addition, with the vast range of occurrences of information in such sources and the dynamic nature of the digital world (e.g., the Internet and the like), computer automated methods and systems are provided to mitigate the likelihood that information becomes stale, is unresolved, is incorrectly interpreted, and the like, while providing timely (e.g., near real-time, and the like) delivery of the information to facilitate various processes and services, such as insurance services, processes and the like. In embodiments, to the extent that a business service, such as providing insurance services, uses sophisticated analytic models of real world conditions when determining if and how much to charge for such services, while remaining competitive, providing clarity of such information in a timely manner (e.g., as requested, such as on-demand, or as conditions change and the like) may be crucial for delivering outcomes from the analytic models that reflect not only a specific set of input conditions (e.g., a request for insurance coverage), but a validated view of real world conditions as represented by the diverse sources of information. In other words, merely having access to a wide range of diverse sources of data, such as social media data, website text, Internet images, and the like is far from sufficient to produce a viable risk index for each individual request for a service. Determining meaning and intent of such information, such as through the use of machine learning and the like, in context of an individual request involves solving problems presented by the presence of such diversity of information itself, problems that without this information could not even be formed, yet alone resolved. Yet further, converting this meaning and intent into a quantified value that is suitable for use in providing a given business service (e.g., insurance and the like) requires application technical methods that facilitate determining, for a given inquiry a quantified value (e.g., a number, range, relative measure, and the like) from such non-quantified information as text, images and the like.

“The methods and systems herein go even further by facilitating automation of selection of which sources of information to harvest. These source of information selection methods and systems may yield significant improvements in computer capabilities, such as performance and the like, by among other things, limiting the number of websites that must be accessed to gather the information, applying only the computing resources at the optimal times for doing so, avoiding overloading computing servers, such as webservers and the like to further reduce congestion and the resultant retries needed for successful access to the websites and other resources controlled thereby, and the like. With the knowledge of the sources to access and the types of information of value to harvest in those sources, improvements in insurance science (e.g., the analytic models and the like mentioned herein and known in insurance technology, and the like) can be achieved while also reducing the computing resources required to do so.

“In embodiments, an insurance data platform is provided that, among other things, uses data from non-traditional, alternative data sources to generate a risk indicator, such as a risk score, from social media and web data for a business, providing unique underwriting insight into the business that cannot be determined using traditional models (such as ones using credit scores). This enables insurers to accurately qualify and price risks, win profitable business, and alleviate the pressure of high losses. References herein to risk scores may refer to one or more measures or indicators of risk, such as calculated based on one or more inputs, such as from one or more alternative data sources or a combination of alternative and traditional data sources, such as providing an overall or aggregated indication of risk, such as associated with a business or individual, a policy element, or the like; however, except where context indicates otherwise, use of the term “risk score” should be understood in connection with various embodiments to encompass any indicator of risk.

“In embodiments, the insurance data platform may provide information that assists with streamlining the insurance application process, streamlining insurance claims processing, rating, underwriting and assessing risk, setting rates and prices for insurance policies, assessing accuracy and truthfulness of information provided by policyholders, detecting fraudulent activities, predicting outcomes (including the likelihood, type and scale of potential loss events), classifying items and activities, and others.

“Methods and systems are disclosed herein for collecting and processing insurance data and may include a crawling system for collecting data from at least one public alternative data site; and an automated data processing system for processing the data from the alternative data site to facilitate an output risk indicator for use in an insurance system. In embodiments, the alternative data site is at least one of an applicant website, a policyholder website, an applicant social media page, a policyholder social media page, an applicant blog, a policyholder blog, a business rating page regarding an applicant business, a business rating page regarding a policyholder business, a product rating page regarding an applicant product, a product rating page regarding a policyholder product, and an online advertisement by an applicant business.

“In embodiments, the risk indicator is produced by combining data from a plurality of the alternative data sites.

“In embodiments, the risk indicator is used for at least one of targeting an applicant for a sales effort, underwriting an insurance decision, pricing an insurance policy, and monitoring a claim.

“In embodiments, the automated data processing system includes a rules engine for applying at least one rule to the collected data to provide at least one risk indicator relating to at least one of an applicant and a policy holder.

“In embodiments, the automated data processing system applies at least one machine learning system to the collected data to provide at least one risk indicator relating to at least one of an applicant and a policy holder.

“In embodiments, the automated data processing system applies at least one hybrid processing system consisting of at least one rules engine and at least one machine learning system to provide at least one risk indicator relating to at least one of an applicant and a policy holder.

“In embodiments, the risk indicator is a business risk score for a defined category of business and wherein the collected data are processed based on the nature of the category. In embodiments, the category of business is at least one of a restaurant business, a retail business, a hotel business, a bar business, a health business, a fitness business, a beauty business, and a spa business.

“In embodiments, the risk indicator is a risk score for a defined category of individual and wherein the collected data are processed based on the nature of the category. In embodiments, the category of individual is at least one of a smoker, a non-smoker, a physically active individual, a disabled individual, an injured individual, an employed individual, and an unemployed individual.

“In embodiments, the risk indicator is a risk score for a defined category of insurance and wherein the collected data are processed based on the nature of the category. In embodiments, the category of insurance is at least one of homeowner’s insurance, commercial general liability insurance, fire insurance, flood insurance, life insurance, property insurance, health insurance, automotive insurance, motorcycle insurance, boat insurance, and libel insurance.”

There is additional summary information. Please visit full patent to read further.”

The claims supplied by the inventors are:

“1. A system for collecting and processing data, the system comprising: a crawling system of a computing device for collecting the data from at least one public alternative data site; and an automated data processing system for processing the data collected by the crawling system from the at least one public alternative data site, wherein the automated data processing system is configured to generate at least one risk indicator for use in an insurance system.

“2. The system of claim 1, wherein the at least one public alternative data site is at least one of an applicant website, a policyholder website, an applicant social media page, a policyholder social media page, an applicant blog, a policyholder blog, a business rating page regarding an applicant business, a business rating page regarding a policyholder business, a product rating page regarding an applicant product, a product rating page regarding a policyholder product, and an online advertisement by an applicant business, and wherein the at least one risk indicator is used by the computing device to support at least one of targeting an applicant for a sales effort, underwriting an insurance decision, pricing an insurance policy, and monitoring a claim.

“3. The system of claim 1, wherein the automated data processing system includes a rules engine, and wherein the automated data processing system automatically applies at least one rule to the data collected by the crawling system to automatically determine the at least one risk indicator relating to at least one of an applicant or a policy holder.

“4. The system of claim 1, wherein the automated data processing system of the computing device includes a machine learning system configured to automatically determine at least one risk indicator relating to at least one of an applicant or a policy holder based on the data collected by the crawling system.

“5. The system of claim 1, wherein the automated data processing system includes at least one hybrid processing system having at least one rules engine and at least one machine learning system that work in concert to determine at least one risk indicator relating to at least one of an applicant or a policy holder.

“6. The system of claim 1, wherein the at least one risk indicator is a business risk score for a defined category of business, and wherein the data collected by the crawling system is processed based on a nature of the defined category of business associated with the data.

“7. The system of claim 6, wherein the defined category of business is at least one of a restaurant business, a retail business, a hotel business, a bar business, a health business, a fitness business, a beauty business, or a spa business.

“8. The system of claim 1, wherein the at least one risk indicator is a risk score for a defined category of individual and wherein the data collected by the crawling system is processed based on a nature of the defined category of individual associated with the data.

“9. The system of claim 8, wherein the defined category of individual is at least one of a smoker, a non-smoker, a physically active individual, a disabled individual, an injured individual, an employed individual, or an unemployed individual.

“10. The system of claim 1, wherein the at least one risk indicator is a risk score for a defined category of insurance and wherein the data collected by the crawling system is processed based on a nature of the defined category of insurance associated with the data.

“11. The system of claim 10, wherein the defined category of insurance is at least one of homeowner’s insurance, commercial general liability insurance, fire insurance, flood insurance, life insurance, property insurance, health insurance, automotive insurance, motorcycle insurance, boat insurance, or libel insurance.

“12. The system of claim 1, further comprising an application programming interface configured to permit at least one of a service, an application or a program to subscribe to the system for collecting and processing data to obtain the at least one risk indicator.

“13. The system of claim 12, wherein the application programming interface is configured to permit a subscriber to subscribe to a stream of risk indicators.

“14. The system of claim 13, wherein the application programming interface automatically pushes a risk indicator to a user based on a presence of a condition, and wherein the condition is an indicator of a change in risk that exceeds a threshold.

“15. The system of claim 1, wherein the data collected by the crawling system is used by the computing device to automatically assess at least one of a likelihood of fraud by an applicant for insurance or a likelihood of fraud by a policyholder.

“16. The system of claim 1, wherein the data collected by the crawling system is used by the computing device to automatically populate an insurance application, and wherein the computing device determines the at least one risk indicator by at least combining data from a plurality of the at least one public alternative data sites.”

For additional information on this patent application, see: Andrews, Geoffrey R.; Drucker, Max Leslie. Systems And Methods For Collecting And Processing Alternative Data Sources For Risk Analysis And Insurance. Filed September 16, 2021 and posted January 6, 2022. Patent URL: https://appft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PG01&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.html&r=1&f=G&l=50&s1=%2220220005125%22.PGNR.&OS=DN/20220005125&RS=DN/20220005125

(Our reports deliver fact-based news of research and discoveries from around the world.)