“ … Through Models and Machine Learning Ally Atkins*, Advisor: Jue Wang † *Department of Neuroscience, †Department of … ”
Abstract
Kidney stone (KS) disease is the most common disease of the urinary tract. Over 1.3 million individuals visit emergency departments (ED) for stone-related symptoms, which is projected to increase. Since most lab tests are misinterpreted, leading to an incorrect diagnosis, diagnostic predictive tools are useful to most accurately calculate an individual's likelihood, or risk, of having KS disease. We investigate which predictive model would yield the highest accuracy and sensitivity by extracting information from a clinical database containing patient data and utilize various machine learning and statistical analysis to analyze the performance of each model. With using confusion matrices, logistic regression, and ROC curves, it has been found that including all domains in a predictive model (demographics, laboratory tests, ICD-9 diagnoses, etc.) demonstrated the highest performance.
“ … who are susceptible to dementia. Christina Vu Advisor: Jue Wang … ”
Abstract
Dementia is a symptom of many neurodegenerative disorders that have a heavy economic burden. In this research, we are building a diagnostic model utilizing a big data approach to analyze clinical records of demographics, diagnoses, vital signs, and other clinical information. We used a large database from one hospital in which thousands of patients were admitted over a span of a decade. We extracted multiple variables which would be relevant for diagnosing patients with dementia and analyzed them through machine learning models.
Encryption and decryption mechanisms, requiring a public algorithm and a secure key, are a way of having a confidential communication over an insecure network, like the Internet. This raises the Key Exchange Problem, informally, the process by which two entities decide on a shared secure key in the unauthenticated network. Password Authenticated Key Exchange (PAKE) protocols attempt to solve the problem by assuming a common secret, such as a shared password, between the parties, which is used to generate the secure key. Our work focuses on such PAKE protocols in an attempt to understand their practical efficiency. Therefore, we implemented certain studied PAKE protocols by applying them to a simple text messaging system. We further tested their practicality through described timing experiments. The Simple Password-based Authentication Key Exchange (SPAKE) protocol and Password Authenticated Key Exchange by Juggling (JPAKE) protocol are two such protocols that we focused on. In working towards an understanding of password-based protocols, as background, this work also reviews certain protocols for Key Exchange within both Private-Key Cryptography and Public-Key Cryptography.
For decades, the United States has committed to economic integration, but Trump's anti-globalization policies have been challenging this commitment and raised concerns and uncertainty about global trade. Hence, we are interested in investigating how that uncertainty affects investor sentiment. This paper uses the S&P 500 index as a proxy for the market and follows the methodology in Bijl et al. (2016) to compute an investor sentiment index by searching company names listed on the S&P 500 index from 2010 to 2019 on Google Trends. Building upon existing literature, this paper proposes a new model that explains the relationship between trade policy uncertainty and investor attention. It is shown that these two variables are positively correlated. Since Google Trends only display the popularity of searches, but cannot reveal the actual sentiment of people who made these searches, this can be interpreted as investors are curious to learn more about companies that they are interested in as the uncertainty increases. Hence, investors are likely to educate themselves in terms of financial knowledge when the forecast is worrisome. This causal flow from TPU to investor attention is also confirmed by a Granger Causality Test.
“ … Abstract Results Conclusion Joseph Palko*, Advisor: Jue Wang * *Department of Mathematics, Union College, … ”
Abstract
Kidney stone disease has become more prevalent through the years, leading to high treatment cost and associated health risks. In this study, we explore a large medical database and machine learning methods to extract features and construct models for diagnosing kidney stone disease. Data of 46,250 patients and 58,976 hospital admissions were extracted and analyzed, including patients' demographic information, diagnoses, vital signs, and laboratory measurements of the blood and urine. We compared the kidney stone (KDS) patients to patients with abdominal and back pain (ABP), patients diagnosed with nephritis, nephrosis, renal sclerosis, chronic kidney disease, or acute and unspecified renal failure (NCA), patients diagnosed with urinary tract infections and other diseases of the kidneys and the uterus (OKU), and patients with other conditions (OTH). We built logistic regression models and random forest models to determine the best prediction outcome. For the KDS vs. ABP group, a logistic regression model using the five variables including age, mean respiratory rate, blood chloride, blood creatinine, and blood CO2 levels from the patients' first lab results gave the best prediction accuracy of 0.699. This model maximized sensitivity with a value of 0.726. For KDS vs. NCA we found that a logistic regression using the Elixhauser score and blood urea nitrogen (BUN) values from the first lab results for patients with first admittance produced the best outcome, with an accuracy of 0.883 and maximized specificity of 0.898. For KDS vs. OKU a logistic regression using the estimated glomerular filtration rate (EGFR) calculated from the average lab values gave the best outcome, with an accuracy of 0.852 and maximized specificity of 0.922. Finally, a logistic regression using age, EGFR, BUN, blood creatinine, and blood CO2 gave the best outcome for KDS vs. OTH, with an accuracy of 0.894 and maximized specificity of 0.903. This research gives the medical field models to potentially use on kidney stone patients. It also provides a steppingstone for researchers to build off if they want to build kidney stone models for a different population of patients.
“ … . Matthew Anderson and Prof. Jue Wang who have constantly been helping me in my work and … whenever … : Prof. Matthew Anderson (CS Dept) & Prof. Jue Wang (Math Dept) i … ”
Abstract
Encryption and decryption mechanisms, requiring a public algorithm and a secure key, are a way of having a confidential communication over an insecure network, like the Internet. This raises the Key Exchange Problem, informally, the problem of two entities deciding on a shared secure key in the unauthenticated network. Password Authenticated Key Exchange (PAKE) protocols attempt to solve the problem by assuming a common secret, such as a shared password, between the parties. This secret is used to generate a secure shared key. Our work focuses on such PAKE protocols in an attempt to understand their practical efficiency. Therefore, we implemented certain studied PAKE protocols by applying them to a simple text messaging system. We further ran described timing experiments to test their practicality.
The Simple Password-based Authentication Key Exchange (SPAKE) and the Password Authenticated Key Exchange by Juggling (JPAKE) are two such protocols that we focus our study on. In working towards an understanding of password-based protocols, as background, this work also reviews certain protocols for Key Exchange within both Private-Key Cryptography and Public-Key Cryptography.
“ … Reserve Currencies ADVISORS - Eshragh Motahar and Jue Wang Throughout most of post World War II period, the … ”
Abstract
Throughout most of post World War II period, the United States dollar has been globally accepted as the dominant reserve currency. This dominance comes with "exorbitant privilege" or special benefits such as not having a balance of payments problem. Therefore, with the shifting of global geopolitical balance of power in the age of Trump, along with the recognition by the IMF of the Chinese renminbi as an international reserve currency in 2015, it is important to understand the modern influence of reserve currencies. We provide an updated study of the status of the dollar, the euro, and in particular, the renminbi as it is an upcoming reserve currency. To do so, we use currency exchange rate data and apply modified workhorse regression models analyzing the co-movement of 149 countries' currencies with each of the five major reserve currencies. We are then able to assign each country's gross domestic product at purchasing power parity to a reserve currency bloc in order to obtain a global sphere of influence for each reserve currency. We find that the United States retains its dominance but faces challenges from the renminbi and the euro in recent years as the international monetary system becomes tri-polar. We analyze particular current events in specific countries and regions to help explain why we see this shifting dynamic.
“ … Department of Mathematics, March 2021. ADVISOR: [Professor Jue Wang ] Kidney stone disease has become more prevalent … ”
Abstract
Kidney stone disease has become more prevalent through the years, leading to high treatment cost and associated health risks. In this study, we explore a large medical database and machine learning methods to extract features and construct models for diagnosing kidney stone disease.
Data of 46,250 patients and 58,976 hospital admissions were extracted and analyzed, including patients' demographic information, diagnoses, vital signs, and laboratory measurements of the blood and urine. We compared the kidney stone (KDS) patients to patients with abdominal and back pain (ABP), patients diagnosed with nephritis, nephrosis, renal sclerosis, chronic kidney disease, or acute and unspecified renal failure (NCA), patients diagnosed with urinary tract infections and other diseases of the kidneys and the uterus (OKU), and patients with other conditions (OTH). We built logistic regression models and random forest models to determine the best prediction outcome.
For the KDS vs. ABP group, a logistic regression model using the five variables including age, mean respiratory rate, blood chloride, blood creatinine, and blood CO2 levels from the patients' first lab results gave the best prediction accuracy of 0.699. This model maximized sensitivity with a value of 0.726. For KDS vs. NCA we found that a logistic regression using the Elixhauser score and blood urea nitrogen (BUN) values from the first lab results for patients with first admittance produced the best outcome, with an accuracy of 0.883 and maximized specificity of 0.898. For KDS vs. OKU a logistic regression using the estimated glomerular filtration rate (EGFR) calculated from the average lab values gave the best outcome, with an accuracy of 0.852 and maximized specificity of 0.922. Finally, a logistic regression using age, EGFR, BUN, blood creatinine, and blood CO2 gave the best outcome for KDS vs. OTH, with an accuracy of 0.894 and maximized specificity of 0.903. This research gives the medical field models to potentially use on kidney stone patients. It also provides a steppingstone for researchers to build off if they want to build kidney stone models for a different population of patients.
“ … this project. I would also like to thank Professor Jue Wang of Math department for helping me learn the … ”
Abstract
Technological dominance and spillovers play important roles in a firm's decision to investment in innovated products. It is intuitive to think that a firm which is technologically ahead will dominate the market for innovated products. However, the question of the spillover advantage a firm gets when they are technologically behind makes the decision to invest in new technology more complex. Therefore, in this paper, I consider the investment in new product and cost of doing research, along with capital and level of technology, to be primary factors affecting a firm's profit. I ask, when is it a good time to invest in new product and when is it appropriate for a firm to allocate more funding for research? I find that firms tend to do more research when they have more market share and invest less when the total capital in the market increases. They tend to invest more when they are technologically ahead and do more research when they are behind. This is a dynamic game because a firm's decision to invest depends not only on its own level of technology but also on the rival firm's level of technology and market share. The presence of technological spillovers also adds onto the dynamics of the game since it discourages the firm, which is technologically ahead, from investing in technological advances.
“ … WoJJ [0S pue ‘peejsawioy pIesag pio @yj Jo juoJy ur pe - Jue [d 2JoM BYSE[y WO} S99I} Aqeq swog *Joxeads B SB 219(]) … ® ur SjUBUL} JO j[oAd1 3y} jJO mo Summos S$IqNOI} JuII-[ jue oy} Ul Sul 1d_Inoj 1o, U0 M3S "H WEIIM … ”
