Swasthya Setu AI – AI-Enabled Healthcare Decision Support Platform

Swasthya Setu AI – AI-Enabled Healthcare Decision Support Platform

AI-Enabled Healthcare Decision Support Platform

Lajari Pawanikar (1), Ojasvi Topale (2), Shivaji Bhoyar (3), Om Kurhade (4), Dr. Komal R. Hole (5)

Department of Computer Science and Engineering (6)

Prof. Ram Meghe Institute Of Technology And Research, Badnera-Amravti, India (7)

ABSTRACT

SwasthyaSetu AI is an AI enabled healthcare platform that aims to simplify the process of selecting and scheduling diagnostic tests, while making it more transparent and affordable.The platform tackles the issue of disparate costs for identical medical tests among diagnostic labs and the challenges patients encounter when comparing laboratory services .The system gathers real-time data from several diagnostic labs using Web Scraping and API Integration. This data includes test costs, ratings, turnaround time (TAT), NABL accreditation, home sample collection availability, distance, and discounts. An AI Recommendation Engine processes this data, taking into account a number of variables instead than just price to suggest the best labs

Additionally, the platform analyzes pricing data and forecasts future price patterns using Machine Learning- based Price Prediction. Users can find the greatest deals by using an automated coupon and discount module. Users can make a one-click online reservation, schedule home sample collection, compare laboratories, and choose the suggested alternative. By combining AI, ML, web technologies, and real-time diagnostic data, SwasthyaSetu AI seeks to enhance healthcare price transparency, lower diagnostic costs, save users’ time, and promote well-informed healthcare decisions.

Keywords:- Web Scraping, API Integration, Diagnostic Test Price Comparison, Price Prediction, Recommendation Engine, Smart Healthcare, NABL Accreditation, Medical Test Booking, Home Sample Collection, Coupon Optimization, Healthcare Price Transparency.

1. INTRODUCTION

   One of the most significant industries that has a direct impact on people’s quality of life is healthcare. Due to the rising incidence of chronic illnesses and the significance of preventive healthcare, the demand for diagnostic services such blood tests, urine tests, MRIs, CT scans, X-rays, ultrasounds, ECGs, and whole body health examinations has dramatically increased in recent years. Early disease identification, diagnosis, and monitoring are made possible by diagnostic tests, which help physicians treat patients

   accurately and promptly[1],[2],[3],[10]. The lack of transparency in diagnostic test cost, however, remains a significant obstacle for patients despite advances in healthcare technology.

   The cost of the same diagnostic test varies significantly amongst labs and diagnostic

   facilities in India. A Complete Blood Count (CBC) test, for example, can cost 350 at one facility and 700 at another in the same city. Similar to this, the cost of full-body checkup packages, thyroid tests, CT scans, MRI scans, and vitamin testing varies greatly depending on the laboratory, region, and current deals. Patients frequently rely on local labs or internet searches because there is no centralized platform to compare these costs, which can result in needless medical costs and poor decision-making[7].

   Before choosing a diagnostic laboratory, patients must consider a number of crucial variables in addition to cost. Laboratory accreditation (such as NABL certification), customer ratings and reviews, report delivery time (Turnaround Time, or TAT), availability of home sample collection, quality of service, distance from the patient’s location, and continuing discounts or promotional offers are some of these criteria[1],[2],[3].

   It takes a lot of effort and can be confusing to compare all these features across several laboratory websites, particularly for elderly people, emergency patients, and people who live in remote places. Because of this, a lot of spatients either choose laboratories without knowing enough about their quality and dependability or overpay for diagnostic services.

   It is now feasible to develop intelligent healthcare solutions that increase accessibility, cost, and transparency because to the quick development of artificial intelligence (AI), machine learning (ML), data analytics, and web technologies. AI is capable of analysing vast amounts of healthcare data, seeing trends, producing precise forecasts, and offering users tailored recommendations [4],[8],[9],[11].

   Healthcare services can become more user-focused and efficient by integrating AI with real-time data collection methods like Web scraping and API integration. The platform incorporates a Machine Learning-based Price

   Prediction Model to analyze historical pricing trends and estimate future price fluctuations for various diagnostic tests. Additionally, an intelligent Recommendation Engine evaluates multiple factorsincluding test price, laboratory reliability, customer reviews, NABL accreditation, turnaround time, distance, and available discountsto recommend the most suitable laboratory for each user[6][12]. The system also has an Automatic Coupon Recommendation Module that searches for and applies the best available promotional offers, allowing users to maximize their savings without having to manually search for discounts. The primary goal of SwasthyaSetu AI is to reduce the cost of diagnostic tests by 2040%, introduce more transparency in healthcare pricing, save the patients valuable time and help them make informed decisions through AI-powered recommendations. The proposed platform combines Artificial Intelligence, Machine Learning, Recommendation Systems, Web Scraping and modern web design technologies to provide an innovative and practical solution to one of the most common challenges faced by healthcare consumers.

   The website collects real-time data from several diagnostic labs, such as test costs, customer reviews, NABL accreditation status, report turnaround time, availability for home sample collection, and current promotional offers. AI algorithms are used to examine this data, which is kept in a central database.

   The Recommendation Engine ensures that consumers receive recommendations that strike a compromise between affordability, quality, dependability, and convenience by ranking laboratories according to a variety of factors rather than just price[10]. Future pricing trends are predicted by a machine learning-based Price Prediction module, and consumers don’t have to seek for the best deals thanks to an automated coupon optimization tool.In order to provide a smooth healthcare experience, the platform also facilitates one-click booking, safe online payments, digital report access, and booking history[11].

   Several stakeholders can profit greatly from SwasthyaSetu AI:

   • Patients benefit from lower diagnostic expenses, clear pricing, easy scheduling, and informed decision- making.

   • Diagnostic Laboratories: Better service utilization, expanded consumer reach, and increased exposure.

   • Healthcare Providers: Enhanced patient satisfaction and better diagnostic service coordination.

   • Healthcare Industry: Encouragement of inexpensive healthcare, healthy competition, and digital change.

   The platform illustrates how intelligent technologies may enhance healthcare accessibility and efficiency while solving real-world issues that millions of patients confront by fusing AI with real-time diagnostic data.

2. LITERATUE SURVEY

   1. Joshua, Thomas, Charles Monwuba, and Onyekas “Design and Implementation of Online Medical Laboratory Diagnostic System” is a web-based laboratory diagnostic system that automates medical testing and lowers errors brought on by human record-keeping is presented in the research article Features including patient registration, appointment scheduling, online payment, report uploading, and result notifications are all included in the system, which was created using the prototyping software development process. The study emphasizes the value of online laboratory services by minimizing in-person interactions and saving patients’ time, particularly during the COVID-19 pandemic.

      This work provides a good foundation for SwasthyaSetu AI. This review describes the system that is designed to deal with a single laboratory only. SwasthyaSetu AI is an extension of this idea that presents AI-based laboratory recommendations, diagnostic test price comparison, cost prediction using Linear Regression, laboratory ratings, turnaround time analysis and multi-laboratory support forming a complete AI-enabled healthcare decision-support platform.

   2. Navyashree C. M. and Manasa C. S.’s “Online Diagnostic Lab Management System” (PHP, MySQL, Apache; SRS methodology) offers role-based booking, digital results, and home sample collection for one lab, supported by UML and ER-based system design, but does not allow patients to compare pricing, ratings, or TAT across labs. SwasthyaSetu AI transforms this single-lab management approach into a comprehensive multi- laboratory decision-support platform with AI-driven recommendations and cross-lab price comparison.

   3. Jaya Venkat, Bhanu Prakash, Naga Sairam Praveen Teja, Pavan Akhil, Gayatri.’s Smart Healthcare Management and Appointment Booking System Django, MySQL, HTML, CSS, Javascript, BootStrap. Three-tier Architecture. This project includes appointment booking with a doctor, video consultation with a doctor, filtering of doctors by specialty, lab test booking with the option of home sample collection, blood donor matching, and a rule- based chatbot. The project uses role-based access for the administrator, physician and patient. The authors discovered that it can decrease waiting times and increase access to better health services, solving the problem of lack of integration and personalization of current systems thru a single digital platform. The SwasthyaSetu AI builds on this foundation especially the home sample collection and appointment scheduling aspects. It adds lab ratings, turnaround time analysis, Linear Regression-based cost projection, cross-lab price comparison, and AI-based recommendations to build a more comprehensive healthcare decision support platform.

   4. Jayatilake and Ganegoda’s paper on machine learning in healthcare decision-making explain the role of

      Machine Learning in healthcare decision-making explaining how the Supervised, Unsupervised, and Deep Learning methods allow the Healthcare Systems to process patients various data such as patients records , medical imaging and other clinical data more efficiently and correctly than traditional approaches.This Paper documents the practical application in image analysis, early diagnosis detection, prediction and personalized treatment planning. The authors note real gains in detecting cancer, diabetes and cardiovascular disease, and say ML was especially useful during COVID-19 when medical staff had to make quick, informed decisions under pressure. In their conclusion they are still cautious: ML certainly improves accuracy and efficiency but should stay a tool to support clinicians, not replace them. And oversight by humans, ethics and patient safety cannot be an afterthought.

   5. Semigran et al conducted an audit of 23 popular symptom-checker apps, testing them against 45 standardized patient cases to see how well they diagnosed conditions and triaged urgency. The results werent great the correct diagnosis appeared first only 34% of the time, although it appeared somewhere in the top 20 suggestions 58% of the time, and the triage advice was appropriate in 57% of cases. The conclusion is that these tools could be a reasonable starting point for patients trying to figure out what’s wrong, but they are not nearly accurate enough to replace an actual doctor. This is important for SwasthyaSetu AI as it gives the reason why the platform does not attempt to diagnose conditions like symptom checkers do, but rather sticks to a narrower, more concrete job comparing labs on price, ratings, TAT and accreditation.

   6. The review-based recommender systems (RBRSs) survey by Hasan et al. in 2025 delves into RBRSs and integrates user reviews, sentiment analysis, and recommendation algorithms to give recommendations that are more believable and personal to the user. The paper discusses various approaches ranging from traditional methods (collaborative filtering, content-based filtering, and hybrid models) to modern techniques (CNNs, RNNs, GNNs, attention mechanisms, and LLMs). It notes that the inclusion of sentiment and user reviews improves the quality of recommendations significantly, while mitigating common problems such as the cold-start problem and sparse data. The problem is that nearly all of this work was developed and tested for online shopping. Swasthya Setu AI employs the same fundamental conceptcombining reviews, ratings, and various choice criteriaand applies it in a way that hasn’t really been done before: assisting patients in selecting a diagnostic lab based on aspects other than simply perusing product listings, such as pricing, quality ratings, and actual reviews.

   7. Araich, Tran, Jung, Horny, and Sadigh (2023) examine

      healthcare price transparency in North America and Europe, finding a common theme: the same service can cost drastically different amounts depending on where you go, and patients have no simple way to compare. And even worse, the price-transparency tools that do exist often are clunky to use and still dont offer people the full picture accurate costs paired with any sense of quality. The authors advice is simple: make these tools easier to use and stop treating price and quality as two separate problems. Thats really the gap Swasthya Setu Al is built to fill, just more focused on a specific, high-friction diagnostic lab tests slice of healthcare – by combining price comparison with laboratory ratings, turnaround time, accreditation and Al- driven recommendations SO patients get transparency and quality context in one place.

   8. In the article published in 2022 by KAIMRC of Saudi Arabia, Abdullah Alanazi explores how machine learning is changing the approach to clinical decision-making as well as healthcare services delivery. He describes the usual advantages such as improved medical images analysis, reduced number of diagnostic mistakes, increased early diagnoses of diseases including

      diabetes, cardiovascular disease, and cancer. Three main types of machine learning are explained in the context of public health management, treatment planning, and disease prediction. Once again, the author mentions the existing challenges, especially related to the need for good quality data and importance of addressing such issues as patient privacy, security, and ethics of using AI technologies. This point of view is consistent with the main idea expressed in the majority of sources reviewed in this research ML has to assist clinicians in making clinical decisions rather than replacing them. SwasthyaSetu AI is based on the same principles, but its application area is much narrower compared to clinical decision-making assisting patients in choosing a reliable and inexpensive diagnostic laboratory with the help of AI recommendations, pricing, and evaluation of laboraory facilities.

   9. In their 2017 work, Jiang et al., discuss the development of AI technology for health applications from machine learning and deep learning to natural language processing being employed in the analysis of patient files, medical imaging, laboratory results, and clinical notes to enhance early diagnosis, treatment planning, and decision-making. They highlight significant advances made in cancer, cardiovascular, and stroke diseases treatment, and emphasize the supportive function of the AI technologies by doctors, rather than AI’s replacement of physicians. The authors do not gloss over such difficult topics as security issues, privacy concerns, and regulatory requirements in relation to the use of AI in clinics. This is the broad perspective of AI application in medicine. SwasthyaSetu AI applies the same AI technologies to address a very particular

   issue laboratory recommendation, diagnostics pricing, cost estimation, and decision making assistance.

3. OBJECTIVES

   1. To obtain the current price comparisons of diagnostic tests from different laboratories using one software that helps users save on their medical expenses by providing the best offers and discounts.

   2. To Provide AI- based Best Deal Recommendation, To recommend the most suitable laboratory based on price, ratings, NABL accreditation, and turnaround time using AI.

   3. To Enable online booking with home sample collection, To allow users to book diagnostic tests online and schedule home sample collection or lab visits conveniently.

   4. To Build trust with AI-Verified reviews and lab ratings, To improve user confidence by providing reliable laboratory ratings ,verified reviews,and quality based recommendations.

4. METHODOLOGY AND WORKFLOW

   The SwasthyaSetu AI workflow describes how a user looks for a diagnostic test, how the system gathers and analyzes laborator data, and how it ultimately suggests the best laboratory. Initially, the user enters basic information like location, age, gender, and name. Next, the user inputs the name and location of the necessary medical test. This aids the system in locating diagnostic labs that are suitable for that specific test and

   fig 1. User Flow

   region. Along with crucial details like pricing, rating, TAT, NABL status, location, and discounts, the system shows the

   laboratories that are available. Instead of visiting many laboratory websites separately, this provides the visitor with all the information they need in one location. Based on factors including cost, quality, turnaround time, accreditation, location, and possible discounts, the customer can compare several laboratories. This makes it simpler to choose which laboratory offers the best value in terms of price, convenience, and quality. The system suggests the best laboratory after comparison. The AI/ML system uses the data it has processed to provide the recommendation, which aids the user in making a better decision. After that, the user can schedule the diagnostic test and choose the suggested laboratory. The user can also arrange for a convenient home sample collection if that option is available. The user completes the secure payment after choosing the test and collection option. The booking confirmation is then given by the system. The user receives the digital diagnostic report via the platform once the diagnostic test is finished

   The system adheres to a technological

   procedure behind the user interface. Web scraping and API integration are the first steps, gathering data in real time from various diagnostic labs.Test price, ratings, TAT, NABL certification, coupons/discounts, home collection availability, and distance are among the data that the system gathers. After that, the data is forwarded for preparation.The gathered information is cleaned, filtered, and merged during data preprocessing so that the AI/ML models may use it

   appropriately.The AI/ML Processing Engine receives the pre-processed data. Price Prediction utilizing Linear Regression, Recommendation Engine, and Coupon Optimization are its three main parts. To evaluate pricing data and forecast prices, the Price Prediction module use Linear Regression. To find an appropriate laboratory, the Recommendation Engine takes into account variables like cost, rating, NABL certification, turnaround time, and distance. The best possible discount is found and automatically applied via the Coupon Optimization module.The Best Laboratory Recommendation Output, which contains crucial data like the best pricing, rating, TAT, NABL status, and reviews, is finally produced by the system.

   1. System Architecture:

      The proposed Swasthya Setu AI architecture consists of five major components:

      1. User Interface

      2. Data Collection Module

      3. Data Preprocessing Module

      4. AI/ML Processing Engine

      5. Laboratory Recommendation and Result Display Module

         The user selects a diagnostic test and specifies a location. The system retrieves relevant laboratory listings from available data sources. The collected information undergoes validation, cleaning, normalization, and duplicate detection before being stored in the database.

         The processed information is passed to the price estimation and recommendation modules. The price estimation module predicts prices only when sufficient historical observations are available. The recommendation engine ranks eligible laboratories using a weighted scoring function. The resulting interface displays relevant laboratory information to help users compare available options.

         Online booking, payment processing, and digital report access are additional platform

         capabilities that must be implemented and tested before being described as operational features.

   2. Data Collection :

      The data collection module gathers diagnostic laboratory information from authorized APIs, participating laboratories, and permitted public web sources.

      The proposed data fields include:

      • Laboratory name and identifier

      • Diagnostic test name and category

      • Listed test price

      • Location and distance

        • Customer rating and review information

      • NABL accreditation status

      • Turnaround time

      • Home sample collection availability

      • Applicable discounts and coupons

      • Data source and collection timestamp

        Each record should retain its source and last verification time. This information helps identify outdated listings and supports traceability.

        The system should comply with applicable website terms, access restrictions, and data-protection requirements. Data availability and update frequency depend on the capabilities and permissions of the selected sources.

   3. Data Preprocessing:

   Data preprocessing is necessary to improve the consistency and reliability of the information used by the recommendation and price estimation modules.

   The preprocessing stage includes:

   Data Cleaning: Invalid, malformed, or incomplete records are identified and flagged for correction or exclusion.

   Duplicate Detection: Duplicate listings are identified using a combination of laboratory identifier, diagnostic test, location, and observation period.

   Missing Value Handling: Missing prices and service attributes are retained as missing rather than incorrectlyconverted to zero. Records lacking essential information may be excluded from particular calculations.

   Data Normalization: Test names, price formats, currencies, location formats, and turnaround-time units are standardized.

   Data Validation: Laboratory details, listed prices, accreditation status, and discount conditions are checked against available sources.

   Timestamp Management: Each record stores the collection time and last verification time to indicate the freshness of the information.

   These preprocessing steps help reduce errors and improve the reliability of the displayed laboratory comparisons.

   LINEAR REGRESSION-BASED PRICE PREDICTION:

   1. Mathematical Formulation

      Linear Regression predicts diagnostic test prices using input features such as test category, laboratory, location, and historical prices.

      P = + x

      Here, P is the predicted price, is the intercept, represents model coefficients, and x represents input features.

   2. Model Training

      The model uses historical pricing data and estimates coefficients by minimizing prediction errors.

      min_ (P P)²

      The dataset should be divided into training, validation, and testing sets. Chronological splitting is preferred for future price prediction.

   3. Model Evaluation

   Model performance is evaluated using:

   • MAE: Average absolute prediction error.

   • RMSE: Penalizes larger prediction errors.

     • R²: Measures the variation in prices explained by the model.

   The results should be compared with a baseline model. Actual performance metrics must be reported only after testing on real data.

   MULTI-CRITERIA RECOMMENDATION ENGINE

   1. Recommendation Approach

      The engine evaluates laboratories using price, ratings, NABL accreditation, turnaround time, distance, discounts, and home sample collection availability. Laboratories that fail mandatory user requirements are excluded before ranking.

   2. Weighted Scoring Function

      S(l) = ws(l) w = 1, w 0

      Here, S(l) is the laboratory score, s(l) is the normalized criterion score, and w is its weight. Higher scores indicate more preferred options. Weights must be documented, and missing information must be handled transparently.

   3. Laboratory Ranking

      Eligible laboratories are ranked in descending order of their final scores. The system displays prices, ratings, accreditation status, turnaround time, distance, discounts, and data verification time.

   4. Recommendation Evaluation

   The engine should be evaluated using data verification rates, constraint violations, and ranking metrics such as Precision@k or NDCG@k when suitable relevance labels are available. Recommendation accuracy should not be claimed without evaluation.

5. Comparison Table

6 .FUTURE SCOPE AND CONCLUSION

The healthcare industry is growing very fast, and because of this, the demand for diagnostic services has also increased. But one big problem is that prices vary a lot from lab to lab, and pricing is not clear. Because there is no single platform to compare, patients have to check many websites, and

still they may end up paying more for the same test. Things like discounts, home sample collection, report delivery time, and lab reliability make the decision even more difficult.

To solve these problems, SwasthyaSetu AI provides a smart online latform where users can compare diagnostic tests, find the best labs, and book tests easily. By using web

scraping, machine learning, AI, and a recommendation engine, the platform gives real-time price comparison, predicts future prices, suggests the best lab based on many quality factors, and automatically applies the best coupons to save money. SwasthyaSetu AI proves that advanced technology has the potential to help overcome challenges in healthcare while making it more transparent and affordable.

As far as future enhancements are concerned, there are a number of improvements which can be incorporated into this website. One such improvement is adding the feature of comparing medicine prices, where people would have the opportunity to compare the price of medicines from different pharmacies and select the cheapest one. In addition to that, the service of appointment booking for doctors and online consultation can also be added.

Another beneficial feature would be integration with health insurance so that the users can assess their insurance coverage and make cashless claims for diagnostics through the same app. Another valuable addition will be the provision of a secured EHR (Electronic Health Records) system for the patients to keep their records safe for their future references.

To improve user experience, an AI-powered health chatbot can be introduced to guide users in choosing the right diagnostic tests and answer general health questions. The platform can also be connected with wearable devices like fitness trackers and smartwatches, so that real-time health data can be used to give personalized suggestions and preventive advice.

Considering Indias language diversity, adding multilingual support is important, so that users from different regions can use the platform easily in their own language. On the technical side, the recommendation engine can be improved using deep learning models to give more accurate and personalized lab suggestions.

Integrating SwasthyaSetu with government schemes related to healthcare such as

Ayushman Bharat will allow users with eligibility to avail free or subsidized diagnostic tests, thus ensuring fair access to healthcare facilities. Thus, the integration of laboratories all over India will ensure the creation of a nationwide laboratory network where users can choose and book tests based on comparisons, making the entire diagnostic ecosystem competitive, transparent, and efficient.

To conclude, it should be said that besides solving existing issues, SwasthyaSetu AI also lays the foundation of a comprehensive AI-enabled healthcare platform. Through providing patients with relevant information, recommendations, and booking options, this system helps ensure a smarter and cost-effective healthcare ecosystem. Through further development and deployment, this platform can become an important tool in the healthcare decision- making process as well as fair access to quality diagnostic facilities.

7. REFERENCES

1. J. V. Joshua, C. K. Thomas, E. T. Charles Monwuba, and O. J. Onyeka, Design and implementation of online medical laboratory diagnostic system, 2023.

2. N. C. M. Navyashree and M. C. S. Manasa, Online diagnostic lab management system, JSS College of Arts, Commerce & Science, 2023.

3. P. J. Venkat, P. B. Prakash, V. N. S. P. Teja, B. P. Akhil, and J. Gayatri, Smart healthcare management and appointment booking system, 2026.

4. S. M. D. A. C. Jayatilake and G. U. Ganegoda, Involvement of machine learning tools in healthcare decision making, Journal of Healthcare Engineering, vol. 2021, Art. no.

   6679512, 2021, doi:

   10.1155/2021/6679512.

5. H. L. Semigran, J. A. Linder, C. Gidengil, and A. Mehrotra, Evaluation of symptom checkers for self diagnosis and triage: Audit study, BMJ, vol. 351, Art. no. p480, 2015, doi: 10.1136/bmj.p480.

6. E. Hasan, M. Rahman, C. Ding, and J. X. Huang, Review-based recommender systems: A survey of approaches, challenges and future perspectives, ACM Computing Surveys, 2025, doi: 10.1145/3742421.

7. H. Araich, J. Tran, J. Jung, M. Horný, and G. Sadigh, Healthcare price transparency in North America and Europe, British Journal of Radiology, vol. 96, no. 1151, Art. no. 20230236, 2023, doi: 10.1259/bjr.20230236.

8. A. Alanazi, Healthcare and opportunities through machine learning, 2022.

9. F. Jiang, Y. Jiang, H. Zhi, Y. Dong, H. Li, S. Ma, Y. Wang, Q. Dong,

   H. Shen, and Y. Wang, Artificial intelligence in healthcare: Past, present and future, Stroke and Vascular Neurology, vol. 2, no. 4,

   pp. 230243, 2017, doi: 10.1136/svn-2017-000101.

10. C. Valenzuela-Núñez, G. Latorre-Núñez, and F. Troncoso-Espinosa, “Smart Medical Appointment Scheduling: Optimization, Machine Learning, and Overbooking to Enhance Resource Utilization,” IEEE Access, vol. 12, pp. 75517562, Jan. 2024, doi: 10.1109/ACCESS.2024.3349953.

11. S. V. Chinta, Z. Wang, A. Palikhe, X. Zhang, A. Kashif, M. A. Smith,

    J. Liu, and W. Zhang, “AI-driven healthcare: A review on ensuring fairness and mitigating bias,” PLOS Digital Health, vol. 4, no. 5, e0000864, May 2025, doi: 10.1371/journal.pdig.0000864.

12. F. Villena, “LaboRecommender: A crazy-easy to use Python-based recommender system for laboratory tests,” arXiv:2105.01209, May 2021.