Entity-level Sentiment Analysis in Contact Center Telephone Conversations Xue-Yong Fu Cheng Chen Md Tahmid Rahman Laskar

2025-05-06 0 0 335.48KB 8 页 10玖币

侵权投诉

Entity-level Sentiment Analysis in Contact Center Telephone

Conversations

Xue-Yong Fu, Cheng Chen, Md Tahmid Rahman Laskar,

Shayna Gardiner, Pooja Hiranandani, Shashi Bhushan TN

Dialpad Canada Inc.

Vancouver, BC, Canada

{xue-yong,cchen,tahmid.rahman}@dialpad.com

{sgardiner,phiranandani,sbhushan}@dialpad.com

Abstract

Entity-level sentiment analysis predicts the

sentiment about entities mentioned in a given

text. It is very useful in a business context

to understand user emotions towards certain

entities, such as products or companies. In

this paper, we demonstrate how we developed

an entity-level sentiment analysis system that

analyzes English telephone conversation tran-

scripts in contact centers to provide business

insight. We present two approaches, one en-

tirely based on the transformer-based Distil-

BERT model, and another that uses a convolu-

tional neural network supplemented with some

heuristic rules.

1 Introduction

Businesses that provide Contact Center as a Ser-

vice (CCaaS) often leverage Artiﬁcial Intelligence

(AI) technologies to transcribe telephone conver-

sations and generate aggregated insights reports

for contact centers across various industry verti-

cals. Customers occasionally make evaluative com-

ments about speciﬁc products or companies during

a customer support call to a contact center. These

comments provide valuable competitive insights to

the business, e.g. positive comments may provide

information useful to a marketing department as it

formulates an advertising campaign while negative

comments may provide valuable insights that can

be used to improve a product or a service. In such

scenarios, building a system that can identify user

sentiments towards entities like product or compa-

nies could be useful.

Though Aspect-Based Sentiment Analysis

(ABSA) (Zhou et al.,2019,2020) that aims to ex-

tract the user sentiment expressed towards a spe-

ciﬁc aspect associated with a given target could

be a possible solution to solve such problems, it

should be noted that ABSA has a key limitation

in this regard. For example, in tasks like competi-

tor analysis where the objective is to understand

the overall user sentiment within a certain period

of time on speciﬁc products or general company

services, ABSA techniques could not be useful as

they provide more ﬁne-grained opinions towards

predeﬁned aspects (e.g. features of a product, ease

of use of a software, or aspects of restaurant experi-

ence) instead of providing a generic user sentiment

towards a speciﬁc entity (Zhou et al.,2019).

While building the Entity-level Sentiment Anal-

ysis (ELSA) system for real-world contact center

use-cases, we observe several key challenges. First

of all, to the best of our knowledge, there are no ex-

isting public datasets available for the entity-level

sentiment analysis task. Meanwhile, this task be-

comes more challenging when the requirement is

to use a dataset constructing from telephone con-

versations since constructing a dataset from speech

transcripts generated from telephone conversations

is non-trivial as telephone transcripts are generated

by automatic speech recognition (ASR) systems

that have their own unique characteristics. For in-

stance, an ASR system may have mistranscription

errors as well as linguistic disﬂuencies (e.g. ﬁller

words) (Fu et al.,2021) that usually occur in a

conversational speech dataset (Malik et al.,2021).

The factors mentioned above make the imple-

mentation of an entity-level sentiment analysis

model very challenging to detect user opinions to-

wards entities that appear in contact center calls. In

this paper, we address the existing limitations be-

hind developing an entity-sentiment model for com-

mercial scenarios in the domain of business tele-

phone conversation data in contact centers. Since

there is no suitable publicly available dataset for

the entity-level sentiment analysis task, we brieﬂy

describe how we sampled and annotated the data

for this task. We then propose two approaches

that leverage neural models for this task (i) one

is based on the DistilBERT (Sanh et al.,2019)

model in which we modify its architecture such

that the model can also be utilized to extract the

arXiv:2210.13401v2 [cs.CL] 26 Oct 2022

opinion term(s) while detecting the sentiment po-

larity towards a named entity; (ii) while the other

approach uses a convolutional neural network (dos

Santos and Gatti,2014) supplemented with some

pre-deﬁned heuristic rules. We compare the ef-

fectiveness of both approaches through extensive

experiments and discuss our ﬁndings to provide

valuable insights for future developments of ELSA

models for real world commercial scenarios.

2 Related Work

Since the entity-level sentiment analysis task is

closely related to aspect-level sentiment analysis, in

this section, we ﬁrst brieﬂy review the aspect-level

sentiment analysis task followed by the entity-level

sentiment analysis task in order to clarify the dis-

tinction between these two tasks while discussing

our rationale behind developing an entity-level sen-

timent analysis model for contact centers.

2.1 Aspect-Based Sentiment Analysis (ABSA)

ABSA aims to classify the sentiment polarity of

aspects of certain objects. Many previous studies

are focused on this research (Sun et al.,2019;Tang

et al.,2016;He et al.,2018;Zhao et al.,2020;Zhou

et al.,2019). A more ﬁne-grained related task is

aspect sentiment triplet extraction (ASTE) (Peng

et al.,2020;Xu et al.,2020), which extracts a triplet

– aspect term, opinion term and sentiment – from

the input. Detection of aspects in both ABSA or

ASTE often relies on implicit lexical or semantic

signs, for instance, the food is too spicy suggests

that this comment is about the taste aspect. This

is different from the entity recognition task where

the goal is to detect the named entities in a given

utterance based on the overall context.

2.2 Entity-level Sentiment Analysis (ELSA)

ELSA aims to predict the sentiment of named enti-

ties in a given text input (Steinberger et al.,2011;

Saif et al.,2014). These named entities are usu-

ally application dependent. One recent work on

ELSA is the work of Luo and Mu (2022), where

they studied entity sentiment in news documents.

Another prominent work on ELSA is the work of

Ding et al. (2018), where an entity-level sentiment

analysis tool was proposed for Github issue com-

ments. Contrary to the above studies that focused

on typed text, our focus is on noisy textual data

(i.e., speech transcripts). Moreover, our proposed

models can infer both entity sentiment and corre-

sponding opinion terms for a better analysis of user

sentiments towards products or companies in busi-

ness telephone conversations in contact centers.

3 Task Description

Let us assume that we have an utterance

w1, w2, ..., wn

containing

words. The goal of

the ELSA task is to identify

opinion words

OW=ow1, ow2, ..., owm

, (where

), and

classify the sentiment of the identiﬁed opinion

words towards the target entity

in the given utter-

ance. In Table 1, we show some examples of the

ELSA task to detect user sentiments towards prod-

ucts and organization type entities. In the ﬁrst two

examples, the customer is directly expressing posi-

tive sentiment about the named entity. For instance,

(i) they say “I love it” indicating “Google” in con-

text, or (ii) they are “very impressed” with “MAC”.

In the third and fourth examples, customers are

expressing negative sentiment about a product or

facet associated with the company, e.g., “He has

a hard time ﬁnding a good yogurt from Walmart”

is a comment about the quality of Walmart’s ser-

vice, not a comment about yogurt. Similarly, in the

fourth example, difﬁculty navigating the Instacart

app is indirectly an indication of negative sentiment

concerning Instacart.

4 Dataset Construction

As noted earlier, there is no publicly available

dataset for the ELSA task. We therefore had to

create and annotate our own dataset. The ﬁrst ma-

jor issue that we observed while constructing a

dataset for ELSA is that the entity-level sentiment

events in our telephone transcripts are very infre-

quent. Hence, random data sampling techniques

might yield an imbalanced dataset where most ut-

terances would not have any positive or negative

sentiments towards an entity. We therefore used

two pre-existing models — a named entity recog-

nition (NER) model based on DistilBERT (Sanh

et al.,2019) that was trained to identify

Organiza-

tion

and

Product

type entities and a convolutional

neural network (CNN) (Krizhevsky et al.,2012;

dos Santos and Gatti,2014;Albawi et al.,2017)

sentiment analysis model — to sample 13000 ut-

terances that contained at least one named entity

and one positive or negative sentiment predicted

by these models. To balance the dataset, we sam-

pled an additional 10000 utterances containing at

least one entity and having no polarized sentiments

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

Entity-levelSentimentAnalysisinContactCenterTelephoneConversationsXue-YongFu,ChengChen,MdTahmidRahmanLaskar,ShaynaGardiner,PoojaHiranandani,ShashiBhushanTNDialpadCanadaInc.Vancouver,BC,Canada{xue-yong,cchen,tahmid.rahman}@dialpad.com{sgardiner,phiranandani,sbhushan}@dialpad.comAbstractEntity-levelse...

展开>> 收起<<

Entity-level Sentiment Analysis in Contact Center Telephone Conversations Xue-Yong Fu Cheng Chen Md Tahmid Rahman Laskar.pdf

共8页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Entity-level Sentiment Analysis in Contact Center Telephone Conversations Xue-Yong Fu Cheng Chen Md Tahmid Rahman Laskar

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: