The Impact of Manipulating the Meaning of Category Labels on Supervised Classification and Inference Learning

Abstract This paper describes our approach for the Chinese Medical named entity recognition(MER) task organized by the 2020 China conference on knowledge graph and semantic computing(CCKS) competition. In this task, we need to identify the entity boundary and category labels of six entities from Chinese electronic medical record(EMR). We construct a hybrid system composed of a semi-supervised noisy label learning model based on adversarial training and a rule postprocessing module. The core idea of the hybrid system is to reduce the impact of data noise by optimizing the model results. Besides, we use post-processing rules to correct three cases of redundant labeling, missing labeling, and wrong labeling in the model prediction results. Our method proposed in this paper achieved strict criteria of 0.9156 and relax criteria of 0.9660 on the final test set, ranking first.

Download Full-text

The impact of imagery-evoking category labels on perceived variety

Seeing and Perceiving ◽

10.1163/187847612x648189 ◽

2012 ◽

Vol 25 (0) ◽

pp. 189

Author(s):

Tamara L. Ansons ◽

Aradhna Krishna ◽

Norbert Schwarz

Keyword(s):

Sensory Information ◽

Sensory Experience ◽

Volcanic Soil ◽

Single Product ◽

Category Labels ◽

Sensory Imagery ◽

The Impact ◽

Perceived Variety

Does sensory imagery influence consumers’ perception of variety for a set of products? We tested this possibility across two studies in which participants received one of three alternate coffee menus where all the coffees were the same but the category labels were varied on how imagery-evocative they were. The less evocative labels (i) were more generic in nature (e.g., ‘Sweet’ or ‘Category A’), whereas the more evocative ones related either (ii) to the sensory experience of coffee (e.g., ‘Sweet Chocolate Flavor’ or ‘Smokey-Sweet Charred Dark Roast’) or (iii) to imagery related to where the coffee was grown (e.g., ‘Rich Volcanic Soil’ or ‘Dark Rich Volcanic Soil’). The labels relating to where the coffee was grown was included as a second control to show that merely increasing imagery does not increase perceived variety; it is increasing the sensory imagery relating to the items that does so. As expected, only category labels that evoked sensory imagery increased consumers’ perception of variety, whereas imagining where the coffee was grown did not enhance perception of variety. This finding extends recent research that shows that the type of sensory information included in an ad alters the perceptions of a product (Elder and Krishna, 2010) by illustrating that the inclusion of sensory information can also alter the perceived variety of a set of products. Thus, the inclusion of sensory information can be used flexibly to alter perceptions of both a single product and a set of choice alternatives.

Download Full-text

Laplacian Embedded Infinite Kernel Model for Semi-Supervised Classification

International Journal of Pattern Recognition and Artificial Intelligence ◽

10.1142/s0218001416500221 ◽

2016 ◽

Vol 30 (10) ◽

pp. 1650022 ◽

Cited By ~ 1

Author(s):

Tao Yang ◽

Dongmei Fu ◽

Chunhong Wu

Keyword(s):

Objective Function ◽

Kernel Function ◽

Supervised Classification ◽

Average Distance ◽

Classification Performance ◽

Support Vector ◽

Multiple Features ◽

Processing Data ◽

Benchmark Datasets ◽

The Impact

Promoted by its convexity and low time complexity, Laplacian embedded support vector regression (LapESVR) model based on manifold regularization (MR) has assumed an important role in semi-supervised classification. Conventionally, the LapESVR model is based on a single kernel function that is intrinsically capable of describing one feature mapping relation only. However, when the data to be processed is from a complex dataset where multiple features of the data are required to be treated, the classification performance using the LapESVR based on a single kernel substantially degrade, indicating that the classification requirement in this case is beyond the capability of the LapESVR. In addition, the processing data is often subject to the impact of abnormal data samples; therefore, in practice assigning a fixed value that is related to the average distance of the processing data as the parameter value of kernel function of the LapESVR is by no means optimal. To solve the problems as mentioned regarding the LapESVR, this paper proposes a Laplacian embedded infinite kernel regression (LapEIKR) model. The proposed model combines the multiple kernels linearly to improve its ability of characterization of the processing data, typical in semi-supervised classification of complex datasets, with multiple features. Further, the parameter setting of the multiple kernels of the LapEIKR model is turned into an optimization problem by formulating a corresponding minimum objective function and an iterative algorithm, and then the values of the settings are facilitated to be obtained by a formulated calculation, assuming the optimal values with respect to the designed objective function. Comparative experiments on the UCI datasets, benchmark datasets and Caltech256 datasets show that the proposed LapEIKR model is improving in terms of adaptivity and efficiency.

Download Full-text

Linguistic Patterns for Code Word Resilient Hate Speech Identification

Sensors ◽

10.3390/s21237859 ◽

2021 ◽

Vol 21 (23) ◽

pp. 7859

Author(s):

Fernando H. Calderón ◽

Namrita Balani ◽

Jherez Taylor ◽

Melvyn Peignon ◽

Yen-Hao Huang ◽

...

Keyword(s):

Supervised Classification ◽

Hate Speech ◽

Automatic Detection ◽

Code Word ◽

Detection Methods ◽

Online Activity ◽

Empirical Results ◽

Linguistic Patterns ◽

The Impact ◽

Speech Identification

The permanent transition to online activity has brought with it a surge in hate speech discourse. This has prompted increased calls for automatic detection methods, most of which currently rely on a dictionary of hate speech words, and supervised classification. This approach often falls short when dealing with newer words and phrases produced by online extremist communities. These code words are used with the aim of evading automatic detection by systems. Code words are frequently used and have benign meanings in regular discourse, for instance, “skypes, googles, bing, yahoos” are all examples of words that have a hidden hate speech meaning. Such overlap presents a challenge to the traditional keyword approach of collecting data that is specific to hate speech. In this work, we first introduced a word embedding model that learns the hidden hate speech meaning of words. With this insight on code words, we developed a classifier that leverages linguistic patterns to reduce the impact of individual words. The proposed method was evaluated across three different datasets to test its generalizability. The empirical results show that the linguistic patterns approach outperforms the baselines and enables further analysis on hate speech expressions.

Download Full-text

Identifikasi dan Estimasi Biomassa Hutan Mangrove dengan Menggunakan Citra Landsat (Studi Kasus : Kabupaten Subang, Jawa Barat)

10.31227/osf.io/2pdtk ◽

2017 ◽

Author(s):

Saddam Chair ◽

Ketut Wikantika ◽

Soni Darmawan

Keyword(s):

Support Vector Machine ◽

Mangrove Forest ◽

Supervised Classification ◽

Research Result ◽

Support Vector ◽

Mangrove Area ◽

Tropical Area ◽

Mate Change ◽

The Impact ◽

Mangrove Biomass

AbstrakHutan mangrove memiliki kandungan hutan terpadat di wilayah tropis. Biomassa berhubungan dengan perubahna iklim dan memiliki peran penting dalam siklus kar- bon. Kerusakan mangrove mengakibatkan bertambahnya kandungan karbon pada at- mosfer. Karena itu diperlukan analisis lahan mangrove dan biomassanya untuk men- gurangi dampak kerusakan mangrove. Penelitian dilakukan dengan metode klasifi- kasi terbimbing untuk mendapatkan luas hutan mangrove Kabupaten Subang pada tahun 1993, 2003 dan 2013. Metode klasifikasi terbimbing yang digunakan adalah Support Vector Machine karena memiliki akurasi terbaik. Estimasi biomassa dil- akukan dengan menggunakan persamaan allometrik dari nilai NDVI mangrove untuk dilihat persebarannya pada Kabupaten Subang. Hasil penelitian menunjukkan penurunan luas hutan mangrove dari tahun 1993 dibandingkan dengan tahun 2003 dan tahun 2013 dibandingkan dengan tahun 2003. Rata-rata biomassa hutan man- grove pada Kabupaten subang pada tahun 1993, 2003 dan 2013 berkisar antara 4-5 ton/hektar.Kata kunci: Mangrove, Biomassa, Landsat, Kabupaten Subang, NDVI. AbstractMangrove is the most carbon-rich forest in tropical area. Biomass is related to cli- mate change and has an important role in carbon cycle. Damage to mangrove can increase the amount of carbon in the atmosfer. It is important to analyze mangrove forest and estimate the biomass to reduce the impact of mangrove forest. This research use supervised classification to get mangrove area in Subang regency from 1993, 2003 and 2013. Support vector machine is used because it has the highest accuracy between the other supervised classification. Allometrik equation that based on NDVI is used to estimate the biomass of mangrove. The research result shows mangrove forest in Subamg regency is decreased from 1993 to 2013. The biomass average in Subang regency is ranged between 4-5 ton/hectare.Keywords: Mangrove, Biomass, Landsat, Subang regency, NDVI.

Download Full-text

Masked Graph Convolutional Network

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2019/565 ◽

2019 ◽

Cited By ~ 1

Author(s):

Liang Yang ◽

Fan Wu ◽

Yingkui Wang ◽

Junhua Gu ◽

Yuanfang Guo

Keyword(s):

Supervised Classification ◽

Label Propagation ◽

Unstructured Data ◽

Classification Methods ◽

Convolutional Network ◽

Node Attributes ◽

Classification Tasks ◽

Node Classification ◽

Supervised Classification Methods ◽

The Impact

Semi-supervised classification is a fundamental technology to process the structured and unstructured data in machine learning field. The traditional attribute-graph based semi-supervised classification methods propagate labels over the graph which is usually constructed from the data features, while the graph convolutional neural networks smooth the node attributes, i.e., propagate the attributes, over the real graph topology. In this paper, they are interpreted from the perspective of propagation, and accordingly categorized into symmetric and asymmetric propagation based methods. From the perspective of propagation, both the traditional and network based methods are propagating certain objects over the graph. However, different from the label propagation, the intuition ``the connected data samples tend to be similar in terms of the attributes", in attribute propagation is only partially valid. Therefore, a masked graph convolution network (Masked GCN) is proposed by only propagating a certain portion of the attributes to the neighbours according to a masking indicator, which is learned for each node by jointly considering the attribute distributions in local neighbourhoods and the impact on the classification results. Extensive experiments on transductive and inductive node classification tasks have demonstrated the superiority of the proposed method.

Download Full-text

1070. Handshake Antimicrobial Stewardship as a Model to Prevent Patient Safety Incidents and Recognize Diagnostic Errors

Open Forum Infectious Diseases ◽

10.1093/ofid/ofz360.934 ◽

2019 ◽

Vol 6 (Supplement_2) ◽

pp. S379-S379

Author(s):

Justin B Searns ◽

Manon Williams ◽

Christine MacBrayne ◽

Ann Wirtz ◽

Sarah Parker ◽

...

Keyword(s):

Patient Safety ◽

Antimicrobial Stewardship ◽

Diagnostic Errors ◽

Diagnostic Error ◽

Care Patient ◽

Patient Harm ◽

Administration Error ◽

Category Labels ◽

The Impact ◽

Patient Safety Incidents

Abstract Background Patient safety incidents (PSIs), such as diagnostic errors, are common events that may lead to significant patient harm. Few studies describe the impact that antimicrobial stewardship programs (ASPs) have preventing PSIs and recognizing diagnostic errors. Handshake Stewardship has emerged as a specific ASP model that involves prospective review of hospital-wide antimicrobial ordering with a compressed “second look” of relevant clinical and historical patient data. In person recommendations are then provided directly to the medical team. The objective of this project was to evaluate the potential impact that Handshake Stewardship has on preventing PSIs and recognizing diagnostic errors. Methods Following Children’s Hospital Colorado (CHCO) ASP’s implementation of the Handshake Stewardship model in October 2013, the CHCO ASP team began prospectively self-labeling interventions as “Great Catches” (GCs). These GCs were defined as any ASP intervention that “notably changed the trajectory of patient care.” Patient charts for all GCs from October 2014 through May 2018 were retrospectively reviewed and each intervention was assigned one or more descriptive category labels including: administration error, de-escalation/escalation of therapy, bug-drug mismatch, inappropriate dose/duration, potential adverse effect, alternative diagnosis, additional testing, prevent hospital admission, and epidemiology alerts. In addition, each intervention was scored using the previously validated “Safer Dx Instrument” to determine which GCs intervened on a potential diagnostic error. Results From October 2014 through May 2018 there were 87,322 admissions to CHCO. Our ASP team intervened on 6,735/87,322 (7.7%) of these admissions. Of these, 174/6,735 (2.6%) were prospectively labeled by ASP providers as GCs, of which 44/174 (25%) resulted in new infectious disease consultations. Conclusion Given the frequency and significance of PSIs including diagnostic error, systems are needed to help recognize and prevent patient harm. The Handshake Stewardship model may help prevent PSIs and recognize diagnostic errors among hospitalized children. Disclosures All authors: No reported disclosures.

Download Full-text

Illegal Oil Bunkering Hotspots, Its Implication on Vegetal Depletion in Delta State, Nigeria

Asian Journal of Advanced Research and Reports ◽

10.9734/ajarr/2020/v13i330311 ◽

2020 ◽

pp. 50-64

Author(s):

Ogoro Mark ◽

Onyeanusi Obianuju Divine ◽

Eze Allen Uche

Keyword(s):

Fresh Water ◽

Oil Spill ◽

Supervised Classification ◽

Oil Spills ◽

Secondary Data ◽

Spatial Trend ◽

Oil Pipeline ◽

Oil Spill Detection ◽

The Impact ◽

Vegetal Cover

The study assessed facilities-based activities of illegal oil bunkering and its spatial trend, hotspots across Delta state. Secondary data was obtained from Landsat imageries of 2013 and 2018, National Oil Spill Monitor and National Oil Spill Detection and Response Agency (NOSDRA). The images were classified using supervised classification method, and the coordinates of illegal bunkering sites were overlaid and analyzed using the Differentiate Weighing Technique to express the magnitude of illegal bunkering activities that occurred across space while the coordinate were imported and overlaid on the administrative map of the study area to enable the appreciation and understanding of the trend in facilities-based activities of illegal bunkers across space. Findings revealed that between the years of 2013 through 2018, 162 oil spills was recorded and were spills recorded as a result of illegal bunkering in Delta state. Oil pipeline accounted for over 50 percent of targeted facilities by the operation of the illegal bunkering. Also, there is a noticeable decrease in the area covered by mangrove and fresh water forest in the tune of 68 and 60 percent respectively. This decrease can be attributed to the impact of spill oil on vegetal cover and health. Thus, the study recommends that communities sensitization programs should be encourage educating the host communities on the extent of self-inflicted impacts on the environment by the activities of locals.

Download Full-text

Supervised Classification of RGB Aerial Imagery to Evaluate the Impact of a Root Rot Disease

Remote Sensing ◽

10.3390/rs10060917 ◽

2018 ◽

Vol 10 (6) ◽

pp. 917 ◽

Cited By ~ 13

Author(s):

Chakradhar Mattupalli ◽

Corey Moffet ◽

Kushendra Shah ◽

Carolyn Young

Keyword(s):

Root Rot ◽

Supervised Classification ◽

Aerial Imagery ◽

Root Rot Disease ◽

Rot Disease ◽

The Impact

Download Full-text

Impact of coastal development on mangrove distribution in Cherating Estuary, Pahang, Malaysia

Malaysian Journal of Fundamental and Applied Sciences ◽

10.11113/mjfas.v15n3.1091 ◽

2019 ◽

Vol 15 (3) ◽

pp. 456-461

Author(s):

Zuhairi Ahmad ◽

Muhammad Luqman Mohamad Suharni ◽

Siti Noor Aifa Taib ◽

Muhammad Shaheed Shammodin

Keyword(s):

Supervised Classification ◽

Coastal Development ◽

The Past ◽

Mangrove Ecosystems ◽

Long Term Monitoring ◽

Tourism Attractions ◽

The Impact ◽

Term Monitoring ◽

Mangrove Degradation

This study examined the impact of various coastal developments on mangrove cover along the Cherating Estuary, Pahang, Malaysia between 1997 and 2016. Series of Landsat imageries at Cherating Estuary have been analyzed using supervised classification. Over 19.99% (26.275 ha) of mangrove cover was degraded along the Cherating estuary from its total area of 131.642 ha during the past 20 years. The expansion of development or urban area can be observed in 1997 (27.93 ha) and in 2016 (111.02 ha), showing an increase by 297.0% in 20 years. We identified several causes of mangrove degradation, including mangrove clearing to commercial or residential area, and aquaculture activities. Cherating is one of the main tourism attractions in Pahang. The pace of coastal development throughout the estuary and the coast area had suppressed the mangroves propagation over the years. Long term monitoring of mangrove ecosystems is essential to ensure the survival and sustainability of mangrove associated biodiversity.

Download Full-text