PROPHENTA

PROPHENTA - Unlocking the medical prophecies

Inspiration

Today the physician across the country are doing everything in their power to keep the patients in good health through early prevention and detection. Although the insufficient doctor to patient ratio leads to frequent medical visits with limited physician time resulting in poor-patient rapport with increased medical cost due to multiple consultations, Numerous diseases having similar symptoms also complicates the accurate diagnosis and timely prediction. This can be solved by building the predictive model PROPHENTA, that aids the physicians.

What it does

Our model is trained with the data acquired from various sources including clinical, EHR and healthcare activity data to make predictions about the future outcomes. After sometime, the Random Forest , SVM algorithms of Machine Learning and Yolov8,Fast RCNN and RNN of Deep Learning is able to make accurate and actionable insights.

Predictive models make assumptions dependent on what has occurred before and what's going on at this point. If new data shows changes in what's going on now, the impact on the future can be recalculated as well.

Our model also forecasts the pre-existing related conditions and the preventive protocols

It analyzes the data and plans a course of treatments that will work best for the patients

Our model also incorporates the OCR Technology, on uploading the patients report the necessary parameters can be scraped from the report and the disease can be predicted.

We have also included the AI specialised MediBot , a care companion to predict both text based as well as the image based diseases.

How we built it

For front-end we have used ReactJs and for Backend we have used Flask and for the ML models we have used the Random Forest , SVM , Decision Tree algorithms and for the DL models we have used Yolov8, Fast RCNN and RNN

Challenges we ran into

Processing large volumes of data was time-consuming and resource intensive.

Addressing the financial implications of using the specialized hardware and clous services for the model development.

Dealing with datasets where one class is significantly more prevalent than others leading to biased model outcomes.

Accomplishments that we're proud of

We have learned lot in these 3 days that we have upgraded our skills in machine learning and deep learning domains and Dealing with the OCR technology , was new to us and we were able to incorporate it successfully.

We are proud that we were able to incorporate the medibot , an AI specialized bot in our model.

What we learned

We have learned the machine learning and the deep learning algorithms, ocr technology, data augmentation and the usage of ensemble models.

What's next for PROPHENTA

Integration of diverse native languages enabling the medibot to engage and assist users across the linguistic barriers We would like to include more disease to our disease prediction system. We would like to utilize the blockchain to securely store the patient data.