AN EXTRACTION METHOD OF TIME-SERIES NUMERICAL DATA FROM ENTERPRISE PRESS RELEASES

2021 ◽  
Author(s):  
Dhairya Vyas

In terms of Machine Learning, the majority of the data can be grouped into four categories: numerical data, category data, time-series data, and text. We use different classifiers for different data properties, such as the Supervised; Unsupervised; and Reinforcement. Each Categorises has classifier we have tested almost all machine learning methods and make analysis among them.


Author(s):  
Abeer K. AL-Mashhadany ◽  
Dalal N. Hamood ◽  
Ahmed T. Sadiq Al-Obaidi ◽  
Waleed K. Al-Mashhsdany

<span id="docs-internal-guid-5dcc170c-7fff-e8e4-10d4-4a07701ca923"><span>Unstructured data becomes challenges because in recent years have observed the ability to gather a massive amount of data from annotated documents. This paper interested with Arabic unstructured text analysis. Manipulating unstructured text and converting it into a form understandable by computer is a high-level aim. An important step to achieve this aim is to understand numerical phrases. This paper aims to extract numerical data from Arabic unstructured text in general. This work attempts to recognize numerical characters phrases, analyze them and then convert them into integer values. The inference engine is based on the Arabic linguistic and morphological rules. The applied method encompasses rules of numerical nouns with Arabic morphological rules, in order to achieve high accurate extraction method. Arithmetic operations are applied to convert the numerical phrase into integer value. The proper operation is determined depending on linguistic and morphological rules. It will be shown that applying Arabic linguistic rules together with arithmetic operations succeeded in extracting numerical data from Arabic unstructured text with high accuracy reaches to 100%.</span></span>


2017 ◽  
Vol 2017 ◽  
pp. 1-10
Author(s):  
Zhihua Li ◽  
Ziyuan Li ◽  
Ning Yu ◽  
Steven Wen

Physiological theories indicate that the deepest impression for time series data with respect to the human visual system is its extreme value. Based on this principle, by researching the strategies of extreme-point-based hierarchy segmentation, the hierarchy-segmentation-based data extraction method for time series, and the ideas of locality outlier, a novel outlier detection model and method for time series are proposed. The presented algorithm intuitively labels an outlier factor to each subsequence in time series such that the visual outlier detection gets relatively direct. The experimental results demonstrate the average advantage of the developed method over the compared methods and the efficient data reduction capability for time series, which indicates the promising performance of the proposed method and its practical application value.


Sign in / Sign up

Export Citation Format

Share Document