PolyAnalyst

{{short description|Artificial intelligence and text mining software}}

{{Infobox software

| name = PolyAnalyst

| logo = PolyAnlayst_logo.png

| logo_alt = A key over a pie chart.

| screenshot =

| caption =

| collapsible =

| developer = Megaputer Intelligence

| released = {{Start date and age|1994}}{{cite journal |title=PolyAnalyst - A Machine Dis covery System Inferring Functional Programs|first=Mikhail V.|last=Kiselev|url=https://www.aaai.org/Papers/Workshops/1994/WS-94-03/WS94-03-021.pdf |publisher=AAAI |access-date=15 March 2021 |language=en|journal=AAAI Technical Report |volume=AAAI-94 Workshop on Knowledge Discovery in Databases|date=1994|issue=WS-94-03|page=237-249}}

| latest_release_version = 6.5

| latest_release_date =

| programming language =

| genre = Data science, artificial intelligence, text mining, predictive analytics

| license =

| website = {{URL|https://www.megaputer.com/}}

}}

PolyAnalyst is a data science software platform developed by Megaputer Intelligence that provides an environment for text mining, data mining, machine learning, and predictive analytics. It is used by Megaputer to build tools with applications to health care, business management, insurance, and other industries. PolyAnalyst has also been used for COVID-19 forecasting and scientific research.

Overview

File:Convolutional-neural-network-polyanalyst-flowchart-example.png node.|alt=Nodes are linked to form a directed acyclic graph in PolyAnalyst's GUI.|402x402px]]

PolyAnalyst's graphical user interface contains nodes that can be linked into a flowchart to perform an analysis. The software provides nodes for data import, data preparation, data visualization, data analysis, and data export.{{Cite news|last=Apicella|first=Mario|date=3 July 2000|title=PolyAnalyst 4.1 digs through data for gold|work=Info World|url=https://books.google.com/books?id=dz0EAAAAMBAJ&q=polyanalyst&pg=PA47|access-date=}}{{Cite journal|last1=Zhang|first1=Qingyu|last2=Segall|first2=Richard S.|date=1 December 2008|title=Web mining: a survey of current research, techniques, and software|url=https://www.worldscientific.com/doi/abs/10.1142/S0219622008003150|journal=International Journal of Information Technology & Decision Making|volume=7|issue=4|pages=683–720|doi=10.1142/S0219622008003150|issn=0219-6220|url-access=subscription}} PolyAnalyst includes features for text clustering, sentiment analysis, extraction of facts, keywords, and entities, and the creation of taxonomies and ontologies. Polyanalyst supports a variety of machine learning algorithms, as well as nodes for the analysis of structured data and the ability to execute code in Python and R.{{Cite journal|last1=Zhang|first1=Qingyu|last2=Segall|first2=Richard S.|date=1 January 2010|title=Review of data, text and web mining software|url=https://doi.org/10.1108/03684921011036835|journal=Kybernetes|volume=39|issue=4|pages=625–655|doi=10.1108/03684921011036835|issn=0368-492X|url-access=subscription}}{{Citation|last1=Zhang|first1=Qingyu|title=Commercial Data Mining Software|url=https://doi.org/10.1007/978-0-387-09823-4_65|work=Data Mining and Knowledge Discovery Handbook|pages=1245–1268|editor-last=Maimon|editor-first=Oded|place=Boston, MA|publisher=Springer US|language=en|bibcode=2010dmak.book.1245Z|doi=10.1007/978-0-387-09823-4_65|isbn=978-0-387-09823-4|access-date=3 October 2020|last2=Segall|first2=Richard S.|date=2009 |editor2-last=Rokach|editor2-first=Lior|url-access=subscription}} PolyAnalyst also acts as a report generator, which allows the result of an analysis to be made viewable by non-analysts.{{Cite web|last=Halper|first=Fern|year=2011|title=Predictive Analytics: The Hurwitz Victory Index Report|url=https://hurwitz.com/dev/wp-content/uploads/2017/09/Hurwitz_Predictive_Analytics_9-27-11.pdf|archive-url=|archive-date=|access-date=28 September 2020|publisher= Hurwitz & Associates}} It uses a client–server model and is licensed under a software as a service model.

Business Applications

= Insurance =

PolyAnalyst was used to build a subrogation prediction tool which determines the likelihood that a claim is subrogatable, and if so, the amount that is expected to be recovered.{{cn|date=March 2024}} The tool works by categorizing insurance claims based on whether or not they meet the criteria that are needed for successful subrogation.{{cn|date=March 2024}} PolyAnalyst is also used to detect insurance fraud.{{cite journal |last1=Wang |first1=John |first2=James G.S. |last2=Yang |title=Data mining techniques for auditing attest function and fraud detection |journal=Journal of Forensic & Investigative Accounting |volume=1 |issue=1 |pages=1–24 |date=2009 |doi= |url=http://web.nacva.com/JFIA/Issues/JFIA-2009-1_9.pdf}}

= Health care =

File:Covid19-confirmedcases_2020-07-02-1.png showing Megaputer's COVID-19 forecast|alt=A heat map of the US|300x300px]]

PolyAnalyst is used by pharmaceutical companies to assist in pharmacovigilance. The software was used to design a tool that matches descriptions of adverse events to their proper MedDRA codes, determines if side effects are serious or non-serious, and to set up cases for ongoing monitoring if needed.{{Cite web|title=Life sciences: Increasing speed-to-insight in pharma|url=https://www.kmworld.com/Articles/Editorial/Features/Life-sciences-Increasing-speed-to-insight-in-pharma-135864.aspx?pageNum=2|access-date=22 September 2020|website=kmworld.com| date=7 January 2020 }} PolyAnalyst has also been applied to discover new uses for existing drugs by text mining ClinicalTrials.gov,{{Cite journal|last1=Su|first1=Eric Wen|last2=Sanger|first2=Todd M.|date=23 March 2017|title=Systematic drug repositioning through mining adverse event data in ClinicalTrials.gov|journal=PeerJ|language=en|volume=5|pages=e3154|doi=10.7717/peerj.3154|issn=2167-8359|pmc=5366063|pmid=28348935 |doi-access=free }} and to forecast the spread of the COVID-19 virus in the United States and Russia.{{Cite web|date=29 April 2020|title=COVID-19: Megaputer provides interactive geo-map to forecast peak of active cases in U.S.|url=https://www.thegeospatial.in/covid-19-megaputer-provides-interactive-geo-mapto-forecast-peak-of-active-cases-in-us|access-date=30 September 2020|website=thegeospatial|language=en}}{{Cite web|title=В России представили модели пика заболеваемости COVID-19 в регионах|url=https://www.rbc.ru/technology_and_media/17/05/2020/5ec0f9779a7947731db85cc9|access-date=24 September 2020|website=РБК|date=17 May 2020 |language=ru}}

= Business management =

PolyAnalyst is used in business management to analyze written customer feedback including product review data, warranty claims, and customer comments.{{cite journal|author=Richard S. Segall |author2=Qingyu Zhang |title=Web Mining of Hotel Customer Survey Data |journal=Systemics, Cybernetics and Informatics |volume=6 |issue=6 |pages=23–29 |citeseerx=10.1.1.455.7659}} In one case, PolyAnalyst was used to build a tool which helped a company monitor its employees' conversations with customers by rating their messages for factors such as professionalism, empathy, and correctness of response. The company reported to Forrester Research that this tool had saved them $11.8 million annually.{{Cite news|last=Evelson|first=Boris|date=10 November 2015|title=Vendor Landscape: Big Data Text Analytics|work=Forrester|access-date=}}

SKIF Cyberia Supercomputer

PolyAnalyst is run on the SKIF Cyberia Supercomputer at Tomsk State University, where it is made available to Russian researchers through the Center for Collective Use (CCU). Researchers at the center use PolyAnalyst to perform scientific research and to management the operations of their universities.{{Cite web|last=редакция|first=Любимая|title=В ТГУ открылся Центр коллективного пользования платформой для аналитики big data|url=https://obzor.city/news/649847---v-tgu-otkrylsja-centr-kollektivnogo-polzovanija-platformoj-dlja-analitiki-big-data|access-date=26 February 2021|website=Томский Обзор|language=ru}} In 2020, researchers at Vyatka State University (in collaboration with the CCU) performed a study in which PolyAnalyst was used to identify and reach out to victims of domestic violence through social media analysis. The researchers scraped the web for messages containing descriptions of abuse, and then classified the type of abuse as physical, psychological, economic, or sexual. They also constructed a chatbot to contact the identified victims of abuse and to refer them to specialists based on the type of abuse described in their messages. The data collected in this study was used to create the first ever Russian-language corpus on domestic violence.{{Cite web|title=Ученые ВятГУ совместно с компанией Мегапьютер Интеллидженс разработали чат-бот для помощи жертвам супружеского насилия - Официальный сайт ВятГУ|url=https://www.vyatsu.ru/internet-gazeta/uchenyie-vyatgu-sovmestno-s-kompaniey-megap-yuter.html|access-date=26 February 2021|website=vyatsu.ru}}{{Cite web|title=Суперкомпьютер помогает находить в интернете жертв домашнего насилия {{!}} iot.ru Новости Интернета вещей|url=https://iot.ru/bezopasnost/superkompyuter-pomogaet-nakhodit-v-internete-zhertv-domashnego-nasiliya|access-date=26 February 2021|website=iot.ru}}