Stochastic Optimization Techniques for Quantification Performance Measures
نویسندگان
چکیده
The estimation of class prevalence, i.e., the fraction of a population that belongs to a certain class, is a very useful tool in data analytics and learning, and finds applications in many domains such as sentiment analysis, epidemiology, etc. For example, in sentiment analysis, the objective is often not to estimate whether a specific text conveys a positive or a negative sentiment, but rather estimate the overall distribution of positive and negative sentiments during an event window. A popular way of performing the above task, often dubbed quantification, is to use supervised learning to train a prevalence estimator from labeled data. Contemporary literature cites several performance measures used to measure the success of such prevalence estimators. In this paper we propose the first online stochastic algorithms for directly optimizing these quantificationspecific performance measures. We also provide algorithms that optimize hybrid performance measures that seek to balance quantification and classification performance. Our algorithms present a significant advancement in the theory of multivariate optimization and we show, by a rigorous theoretical analysis, that they exhibit optimal convergence. We also report extensive experiments on benchmark and real data sets which demonstrate that our methods significantly outperform existing optimization techniques used for these performance measures.
منابع مشابه
A stochastic version analysis of an M/G/1 retrial queue with Bernoulli schedule
In this work, we derive insensitive bounds for various performance measures of a single-server retrial queue with generally distributed inter-retrial times and Bernoulli schedule, under the special assumption that only the customer at the head of the orbit queue (i.e., a FCFS discipline governing the flow from the orbit to the server) is allowed to occupy the server. The method...
متن کاملUtilization of Soft Computing for Evaluating the Performance of Stone Sawing Machines, Iranian Quarries
The escalating construction industry has led to a drastic increase in the dimension stone demand in the construction, mining and industry sectors. Assessment and investigation of mining projects and stone processing plants such as sawing machines is necessary to manage and respond to the sawing performance; hence, the soft computing techniques were considered as a challenging task due to stocha...
متن کاملIntroducing a Relational Network DEA Model with Stochastic Intermediate measures for Portfolio Optimization
متن کامل
Robust Stochastic Design of Linear Controlled Systems for Performance Optimization
This study discusses a robust controller synthesis methodology for linear, time invariant systems, under probabilistic parameter uncertainty. Optimization of probabilistic performance robustness for H2 and multi-objective H2 measures is investigated, as well as for performance measures based on first-passage system reliability. The control optimization approaches proposed here exploit recent ad...
متن کاملOptimizing Non-decomposable Performance Measures: A Tale of Two Classes
Modern classification problems frequently present mild to severe label imbalance as well as specific requirements on classification characteristics, and require optimizing performance measures that are non-decomposable over the dataset, such as F-measure. Such measures have spurred much interest and pose specific challenges to learning algorithms since their non-additive nature precludes a dire...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- CoRR
دوره abs/1605.04135 شماره
صفحات -
تاریخ انتشار 2016