DESIGN AND IMPLEMENTATION OF AN APPLICATION FOR MEDICAL DECISION SUPPORT.

Reading multiple product reviews can be tedious, and concluding whether or not consumers liked a product is complicated, so it is necessary to implement a tool that analyzes all reviews of a product and determines their polarity. The foregoing in order to streamline and improve decision-making about a product by the interested parties, as well as the client-company relationship, evaluating the reviews under the same criteria. During the development of the project, the strategy was developed and implemented using Machine learning and Data mining techniques to solve the problem posed. As a result, a model was implemented through a data set, then web scrapping was applied to the Amazon website, a recognized E-commerce, in order to extract the reviews of a given product, the reviews of this product were displayed. through Python libraries to later be processed and thus carry out a sentiment analysis. The above concluded the polarity of a given product making use of machine learning and data mining techniques.

Authors

ADVISOR

Autores

Imitola, Jesús Nieto, Miguel

Utria, Jhan

TUTOR Nieto, Wilson

Ingenier A De Sistemas Y Computaci N

APLICACIÓN DE TÉCNICAS DE INTELIGENCIA ARTIFICIAL PARA EL DIAGNÓSTICO ASISTIDO DE GLAUCOMA.

El fin de este proyecto se centra en la implementación de una solución software que mediante el uso de una red neuronal convolucional permita asistir el diagnóstico de glaucoma tomando como entrada un fondo ocular. Debido a la problemática que representa el glaucoma para la salud ocular en la sociedad, es indispensable el diagnóstico rápido y confiable de esta enfermedad con el fin de tomar acciones ágiles de mitigación y/o reducción que ataquen la pérdida de la visión que causa esta afección. Dentro del proyecto, se busca en general diseñar, implementar y validar un prototipo que permita reducir tiempos en el proceso de diagnóstico, esto, es llevado a cabo siguiendo la metodología de desarrollo CRISP-DM, donde a grandes rasgos el proceso seguido es: el entendimiento del problema a través de una revisión sistemática de la literatura, recopilando la información de estrategias e información ya existentes para la agilización del diagnóstico; la obtención de los datos de alguna fuente, que en este caso se trata de un dataset tomado de la plataforma Kaggle; el entendimiento y preparación de estos datos, donde se busca entender y dar la estructura correcta a los datos para los próximos pasos; el modelado de la solución, donde se realiza el diseño que va a seguir la solución; la implementación, realizada sobre Python utilizando Keras y TensorFlow y por último la validación, donde comprobamos mediante un proceso de validación cruzada que los objetivos planteados hayan sido cumplidos.

The aim of this project is focused on the implementation of a software solution that through the use of a convolutional neural network allows to assist the diagnosis of glaucoma taking as input an ocular fundus. Due to the problems that glaucoma represents for eye health in society, it is essential to have a fast and reliable diagnosis of this disease in order to take agile actions to mitigate and/or reduce the loss of vision caused by this condition. Within the project, the overall objective is to design, implement and validate a prototype that allows to reduce time in the diagnostic process, this is carried out following the CRISP-DM development methodology, where the process followed is broadly: understanding the problem through a systematic review of the literature, gathering information from existing strategies and information to speed up the diagnosis; obtaining data from a source, which in this case is a dataset taken from the Kaggle platform; the understanding and preparation of this data, where we seek to understand and give the correct structure to the data for the next steps; the modeling of the solution, where the design that the solution will follow is made; the implementation, performed on Python using Keras and TensorFlow and finally the validation, where we check through a process of cross-validation that the objectives have been met.

Authors

Imitola, Jesús Nieto, Miguel Utria, Jhan

ADVISOR Nieto, Wilson

Autores

Medina Barrios, Ricardo

José

Walton Romero, Geraldine

Patricia

Castillo Rodado, Gabriel

TUTOR

Nieto Bernal, Wilson

INGENIERÍA DE SISTEMAS Y COMPUTACIÓN