Model Loss and Distribution Analysis of Regression Problems in Machine Learning

Nan, Yun; Zheng, Zeyu; Wang, Tianran

doi:10.1145/3318299.3318367

Cited by 35 publications

(56 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since the cannabinoids were not normally distributed (Fig. 1 A), a fundamental assumption for machine learning algorithms and regression statistics 58 , 59 , it can have the tendency to skew the slope of the regression curve, which may be the reason why higher concentrated cannabinoids are underpredicting. Another limitation observed through this model building process was found for limit level cannabinoid concentrations.…”

Section: Discussionmentioning

confidence: 99%

Using a global diversity panel of Cannabis sativa L. to develop a near InfraRed-based chemometric application for cannabinoid quantification

Gloerfelt-Tarp

Hewavitharana

Mieog

et al. 2023

Sci Rep

View full text Add to dashboard Cite

C. sativa has gained renewed interest as a cash crop for food, fibre and medicinal markets. Irrespective of the final product, rigorous quantitative testing for cannabinoids, the regulated biologically active constituents of C. sativa, is a legal prerequisite across the supply chains. Currently, the medicinal cannabis and industrial hemp industries depend on costly chromatographic analysis for cannabinoid quantification, limiting production, research and development. Combined with chemometrics, Near-InfraRed spectroscopy (NIRS) has potential as a rapid, accurate and economical alternative method for cannabinoid analysis. Using chromatographic data on 12 therapeutically relevant cannabinoids together with spectral output from a diffuse reflectance NIRS device, predictive chemometric models were built for major and minor cannabinoids using dried, homogenised C. sativa inflorescences from a diverse panel of 84 accessions. Coefficients of determination (r2) of the validation models for 10 of the 12 cannabinoids ranged from 0.8 to 0.95, with models for major cannabinoids showing best performance. NIRS was able to discriminate between neutral and acidic forms of cannabinoids as well as between C3-alkyl and C5-alkyl cannabinoids. The results show that NIRS, when used in conjunction with chemometrics, is a promising method to quantify cannabinoids in raw materials with good predictive results.

show abstract

Section: Discussionmentioning

confidence: 99%

Using a global diversity panel of Cannabis sativa L. to develop a near InfraRed-based chemometric application for cannabinoid quantification

Gloerfelt-Tarp

Hewavitharana

Mieog

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In addition, sample complexities of 𝑄-learning variants with (linear) function approximations have been established in several recent papers. Yang and Wang (2019) assumed a linear MDP framework (see the end of Section 2.1) and the authors provided a V-sample complexity ( 𝑑 (1−𝛾) 3 𝜀 2 poly(log 𝛿 −1 )) with 𝑑 denoting the dimension of the features. This result matches the information-theoretic lower bound up to log(⋅) factors.…”

Section: |||mentioning

confidence: 99%

Recent advances in reinforcement learning in finance

Hambly

Yang

2023

Mathematical Finance

View full text Add to dashboard Cite

The rapid changes in the finance industry due to the increasing amount of data have revolutionized the techniques on data processing and data analysis and brought new theoretical and computational challenges. In contrast to classical stochastic control theory and other analytical approaches for solving financial decision‐making problems that heavily reply on model assumptions, new developments from reinforcement learning (RL) are able to make full use of the large amount of financial data with fewer model assumptions and to improve decisions in complex financial environments. This survey paper aims to review the recent developments and use of RL approaches in finance. We give an introduction to Markov decision processes, which is the setting for many of the commonly used RL approaches. Various algorithms are then introduced with a focus on value‐ and policy‐based methods that do not require any model assumptions. Connections are made with neural networks to extend the framework to encompass deep RL algorithms. We then discuss in detail the application of these RL algorithms in a variety of decision‐making problems in finance, including optimal execution, portfolio optimization, option pricing and hedging, market making, smart order routing, and robo‐advising. Our survey concludes by pointing out a few possible future directions for research.

show abstract

“…In addition, there are two targets, SBP and DBP, which are more akin to a www.nature.com/scientificdata www.nature.com/scientificdata/ multi-task or multi-output regression problem 59 . Finally, the distributions of SBP and DBP are often skewed, as extreme BP is much rarer, which makes it an imbalance regression problem 60 .…”

Section: • Ppg Cycle Identification and Bpm Limitationmentioning

confidence: 99%

A benchmark for machine-learning based non-invasive blood pressure estimation using photoplethysmogram

2023

View full text Add to dashboard Cite

Blood Pressure (BP) is an important cardiovascular health indicator. BP is usually monitored non-invasively with a cuff-based device, which can be bulky and inconvenient. Thus, continuous and portable BP monitoring devices, such as those based on a photoplethysmography (PPG) waveform, are desirable. In particular, Machine Learning (ML) based BP estimation approaches have gained considerable attention as they have the potential to estimate intermittent or continuous BP with only a single PPG measurement. Over the last few years, many ML-based BP estimation approaches have been proposed with no agreement on their modeling methodology. To ease the model comparison, we designed a benchmark with four open datasets with shared preprocessing, the right validation strategy avoiding information shift and leak, and standard evaluation metrics. We also adapted Mean Absolute Scaled Error (MASE) to improve the interpretability of model evaluation, especially across different BP datasets. The proposed benchmark comes with open datasets and codes. We showcase its effectiveness by comparing 11 ML-based approaches of three different categories.

show abstract

Model Loss and Distribution Analysis of Regression Problems in Machine Learning

Cited by 35 publications

References 25 publications

Using a global diversity panel of Cannabis sativa L. to develop a near InfraRed-based chemometric application for cannabinoid quantification

Using a global diversity panel of Cannabis sativa L. to develop a near InfraRed-based chemometric application for cannabinoid quantification

Recent advances in reinforcement learning in finance

A benchmark for machine-learning based non-invasive blood pressure estimation using photoplethysmogram

Contact Info

Product

Resources

About