CNN-LSTM- SCSO: An hybrid approach for Forest Fire Prediction

Project Abstract: The forest fire disaster brings forth significant consequences, encompassing threats to both the flora and fauna environment, along with the overarching challenge of global warming. Integrating various Machine Learning methodologies stands out as a crucial approach for monitoring the propagation of fires within forested regions. By doing so, it becomes possible to avert hazardous situations, curbing the potential for environmental and property losses. Consequently, the ability to accurately identify the progression of fire, particularly its impact on burned areas, has gained paramount importance in mitigating the catastrophic aftermath of forest fires. Current research endeavors have been directed towards enhancing the precision of monitoring systems. In this pursuit, a novel approach was adopted: the fusion of Convolutional Neural Networks (CNN), Long Short-Term Memory networks (LSTM), and Sand-cat Swarm Optimization (SCSO). This hybrid methodology, denoted as CNN-LSTM-SCSO, was harnessed for predicting forest fires. Notably, this approach exhibited superior efficacy when compared to conventional methods, marking a significant stride towards proactive fire management and prevention

PanCan Stage Identifier

A Transfer Learning Based Multi- Tumor Classification Framework Approach For Pancreatic Tumor Stage and Substage Classification

Project Abstract:Pancreatic cancer, a leading cause of over 50,000 deaths annually, suffers from poor prognosis mainly due to late detection. Existing diagnostic hurdles, such as difficulty in distinguishing tumors and size variability, call for improved detection methods. This research pioneers the Pan Can Stage Tracker tool, focusing on sub-staging, providing detailed insights into disease progression. Unlike traditional methods hindered by the pancreas's deep location, sub- staging offers precise assessments, aiding in accurate prognostic predictions and treatment decisions. The methodology involves machine learning, leveraging preprocessing of abdominal CT images, transfer learning for tumor classification, and pancreas segmentation through LSTUN segmentation. Feature extraction includes radiomic features, higher-order statistics, and GLCM statistics, with optimal feature selection and data augmentation techniques applied. Importantly, this approach shows promise for repurposing in other cancer types for preprocessing and tumor differentiation. Additionally, the utilization of KCPC in abdominal region mapping enables identification of both stages and substages with an impressive accuracy of over 98%, potentially escalating the 5-year survival rate from 3% to 44% through early Detection.

Shelter Connect Mobile Application

Developed a fully functioning application that allows local users to view nearby shelters and what products they accept or need; used Javascript and x code to create the app.

Included features like a resource locator, AI integrated chat bot, and various health and crisis lines to help the homelessness society.

PedAppl Mobile Application

Developed a fully functioning application that allows users to report local street faults; today, many faulty streets, sidewalks, broken traffic/pedestrian crossing lights that should be reported are now made a lot easier using PedAppl as requests can be sent to the local mayor directly.

Programmed this app using Swift and Xcode Apple App Developer.

Royal Family Kids Camp, Volunteer, 65+ hours, (10th-11th)

Supported young children by volunteering as a full time camp counsler and assisting children in various outdoor activities such as rowing, swimming, ziplining, arts and crafts, etc… Through this organization for children in foster care

Cedar Sinai Park, Volunteer, 8+ hours, (9th-11th)

Worked in this assisted living facility and supported residents by helping the residents move and managing games such as beach ball and bingo

Blanchet House, Volunteer, 10 hours, (9th-10th)

Served food in downtown Portland for the homeless Oregon Food Bank

JET USA

Worked as a tutor under JET Oregon, a spiritual program located in Portland

Worked for 4 months, 3 hours/week for 5+ students.Spiritually involved in local community service events such as collecting recycling and compost/ creating food for homeless

Oregon Museum of Science and Technology (OMSI) Teen Science Alliance, Volunteer / Leader, (9th-11th) (170+ hours)

Created educational videos on topics including photosynthesis for kids to learn at OMSI Demonstrated leadership and commitment not only worked in the physics lab but also taking on managerial responsibilities.

Bit by Bit Coding, Marketing Volunteer, (9th-11th)

Created social media posts weekly/biweekly for a company dedicated to helping students learn more about coding languages Helped renovate/launch Bit by Bit’s newest project, Form and Function, which is a branch of the company that teaches middle schoolers introductory coding/tech

STEM Like a Girl Workshop Volunteer, 9th-10th

Assisted at monthly STEM workshops as the design challenge mentor Workshops consisted of teaching 1 hour javascript game creation through code.org

PSU Center for Entrepreneurship, Workshop Volunteer, (9th-11th)

Worked for 4 months, 3 hours/week for 5+ students.Spiritually involved in local community service events such as collecting recycling and compost/ creating food for homeless

Women in Science Blog, Contributor, (9th-11th)

Dedicated my blog posts to amplifying the voices of women in science. Conducted interviews with accomplished female scientists and engaged in insightful discussions to write engaging narratives about their experiences. Leveraged my writing to shed light on the experiences, achievements, and contributions of these women to their respective fields, inspiring readers to pursue careers in STEM.

TechGirlz Workshop, Volunteer, (9th-11th)

Assisted with the instruction of 2 web design workshops and 1 “Intro to Coding” workshop through the TechGirlz organization 45 minute workshops

UPchieve, Tutor, (9th-11th)

Tutored students grades 5-8 in STEM subjects including biology, all levels of math until calculus Tutored 4 Students weekly for 1 hour sessions.