An agent-based approach for recommending cultural tours

doi:10.1016/j.patrec.2020.01.005

Pattern Recognition Letters

Volume 131, March 2020, Pages 341-347

https://doi.org/10.1016/j.patrec.2020.01.005 Get rights and content

Highlights

•
In the Big Data era recommender systems have become essential to provide intelligent browsing.
•
We provide a model based on Multi Agent Systems to model and forecast users behaviors.
•
Recommendation and Model Checking are used in order to provide plans for tours.
•
Counterexample-driven planning and recommendation allow for fast tours management.

Abstract

In the Cultural Heritage domain, the introduction of intelligent systems for planning of cultural visit was very appealing. Indeed when people decide to visit museums or archaeological sites, they usually would like to organize and schedule their time in order to fulfill some requirements and to match, as far as possible, their preferences and needs. In this work, we propose a novel methodology integrating recommendation facilities with agent-based planning techniques in order to implement a planner of routes within cultural sites as museums. In particular, the introduced methodology exploits from one hand a user-centered recommendation strategy to suggest the most suitable cultural items with respect to user’s preferences and, from the other one, it leverages multi-agents planning methods to generate the related routes consisting of the sequence of steps necessary to reach precise cultural goals depending on the context, by means of a state space exploration. We present a case of study for the proposed methodology and we describe some experimental results on system efficiency.

Introduction

In the age of Big Data recommender systems have become essential to provide intelligent browsing of more and more large collections of items, thus supporting users to effectively and efficiently discover “what they need” within the ocean of digital information.

Cultural Heritage undoubtedly represents one of the main application domains that can benefit from the recommendation facilities [10], [11], [13], [19].

In fact, cultural exhibitions are rapidly moving from an old vision where static information are proposed to users, to a novel one leveraging personalized services able to match visitors’ personal interests and behaviors. The final goal is to provide a “user-centered information dialog” between a cultural space and its visitors [9].

Generally, cultural digital contents come from distributed and heterogeneous data sources such as digital libraries, archives of cultural foundations, multimedia art collections, web encyclopedias, social media networks (like in [5], [15]) and so on.

Indeed when people decide to visit museums or archaeological sites, they usually would like to organize and schedule their time in order to fulfill some requirements, preferences and needs. As an example, they could prefer to see only paintings by a given author or belonging to the same artistic genre, often according to several cultural paths that they have in mind before starting the visit.

Furthermore, sometimes and for particular kinds of museums or archaeological ruins, the planned tours can take very long time to be accomplished and it is not unusual that visitors have to leave before the tour is effectively completed, even if tracking systems are used in order to better manage users’ routes [1]. In addition, a wrong scheduling can overload some areas of cultural sites that have to be managed by using proper queuing policies because of safety and security reasons [4], [12]. In this last case, a re-plan of visitor routes is necessary.

In this work, we propose a novel methodology integrating recommendation facilities with agent-based planning techniques in order to implement a planner for cultural routes within cultural sites and museums.

A case of study for the proposed methodology is reported and some preliminary experimental results on system efficiency were obtained to validate the approach.

Section snippets

Recommender systems for cultural heritage

As well known, recommender systems help people in retrieving information that match their preferences by recommending products or services from a large number of candidates, supporting decisions in various contexts. Just to make some examples, they can suggest what items to buy, which photo or movie to watch, which music to listen, what travels to do, who they can invite to their social network, or even which artwork could be interesting within an art collection [8].

In Content-Based Filtering

Recommendation

Recommendation is easily modeled as an information filtering problem. Here, the objective is to reduce and rank the vast amount of multimedia assets from the different data sources that can be associated to a cultural artifact (e.g., images and textual description related to a given picture within a museum), thus simplifying the agent task of searching the most suitable objects for visitors.

Formally, a recommender system deals with a set of users $U = {u_{1} \dots, u_{m}}$ and a set of items $O = {o_{1}, \dots, o_{n}}$ . For

The recommendation strategy

The basic idea behind the adopted strategy is that when a user is browsing a multimedia art collection in a real o virtual environment, the recommender system: (i) selects a set of useful candidate items on the base of user actual needs and preferences (pre-filtering stage); (ii) opportunely assigns to these items a rank, previously computed exploiting items’ intrinsic features, users’ past behaviors, and also leveraging users’ opinions and feedback (ranking stage); (iii) dynamically, when a

Case studies

The case study we propose, consists in a scenario of a museum, populated by a group of tourists, who want to visit objects of interest inside it. Tourists have their own personal goals. We can have tourists interested in visiting the entire museum, or only particular areas (e.g. Ancient History or Renaissance).

The Environment (World) provides the description of all the areas of the museum and information regarding the entry and exits points of each museum room. It is assumed that the museum is

Conclusions and future works

In the domain of cultural heritage, the introduction of a system for visits planning is very appealing. When people go to visit museums or archaeological sites, they usually want to schedule their time in order to fulfill some requirements, like they would see first paintings from the same period or region. The introduction of a tool able to adapt tours to user expectation and time requirements can noticeably improve users’ experiences and satisfaction. Furthermore, sometimes, tours can take

CRediT authorship contribution statement

Flora Amato: Conceptualization, Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Francesco Moscato: Conceptualization, Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Vincenzo Moscato: Conceptualization, Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Francesco Pascale: Conceptualization, Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Antonio

Declaration of Competing Interest

None.

References (24)

J.A. Baier et al.
A heuristic search approach to planning with temporally extended preferences
Artif. Intell.
(2009)
J. Bobadilla et al.
Recommender systems survey
Knowl.-Based Syst.
(2013)
A. Castiglione et al.
Chis: a big data infrastructure to manage digital cultural items
Future Gener. Comput. Syst.
(2018)
F. Colace et al.
A collaborative user-centered framework for recommending items in online social networks
Comput. Hum. Behav.
(2015)
R. Zhou et al.
Breadth-first heuristic search
Artif. Intell.
(2006)
I. Ahmed, A. Ahmad, F. Piccialli, A. Sangaiah, G. Jeon, A robust features-based person tracker for overhead views in...
M. Albanese et al.
A multimedia recommender integrating object features and user behavior
Multimed. Tools Appl.
(2010)
M. Albanese et al.
A multimedia recommender system
ACM Trans. Internet Technol.
(2013)
F. Amato et al.
Snops: a smart environment for cultural heritage applications
Proceedings of the Twelfth International Workshop on Web Information and Data Management
(2012)
F. Amato, V. Moscato, A. Picariello, F. Piccialli, G. Sperl, Centrality in heterogeneous social networks for Lurkers...

I. Bartolini et al.

Recommending multimedia visiting paths in cultural heritage applications

Multimed. Tools Appl.

(2016)

J.P. Bowen et al.

Personalization and the web from a museum perspective

Museums and the Web

(2004)

Cited by (26)

An accuracy-enhanced group recommendation approach based on DEMATEL
2023, Pattern Recognition Letters
Group recommendations aim to suggest items to a group of users based on their preferences. Many group recommendations often consider various factors to calculate the influence of each group member and then assign weights to complete the group recommendation, aiming to maximize group member satisfaction. However, most group recommendations tend to focus more on the calculation process of group members’ influence while ignoring the process of assigning their weights, which may lead to low group members’ satisfaction. To solve this problem, a novel group recommendation approach called GroupRec_D is proposed to assign weights of users scientifically and reasonably based on data mining and DEMATEL technique. To demonstrate the availability and effectiveness of GroupRec_D, we conduct extensive experiments on the MovieLens 100k dataset and use three evaluation metrics including GSM, RECALL, and nDCG to evaluate the approach. Experimental results demonstrate that GroupRec_D outperforms other comparison approaches.
App usage on-the-move: Context- and commute-aware next app prediction
2022, Pervasive and Mobile Computing
Citation Excerpt :
People spend a large proportion of time commuting for multiple purposes. Studies have revealed that people often engage with their mobile phones to carry out different activities during their commutes [21–23]. However, studies show that engagement with mobile phone activities can cause distraction and compromise safety under certain circumstances such as driving [24–26].
The proliferation of digital devices and connectivity enables people to work anywhere, anytime, even while they are on the move. While mobile applications have become pervasive, an excessive amount of mobile applications have been installed on mobile devices. Nowadays, commuting takes a large proportion of daily human life, but studies show that searching for the desired apps while commuting can decrease productivity significantly and sometimes even cause safety issues. Although app usage behaviour has been studied for general situations, little to no study considers the commuting context as vital information. Existing models for app usage prediction cannot be easily generalised across all commuting contexts due to: (1) continuous change in user locations; and (2) limitation of necessary contextual information (i.e., lack of knowledge to identify which contextual information is necessary for different commuting situations. We aim to address these challenges by extracting essential contextual information for on-commute app usage prediction. Using the extracted features, we propose AppUsageOTM, a practical statistical machine learning framework to predict both destination amenity and utilise the inferred destination to contextualise the app usage prediction with travelling purposes as crucial information. We evaluate our framework in terms of accuracy, which shows the feasibility of our work. Using a real-world mobile and app usage behaviour dataset with more than 12,495 trajectory records and more than 1046 mobile applications logged, AppUsageOTM significantly outperformed all baseline models, achieving $A c c u r a c y @ k$ 46.4%@1, 66.4%@5, and 75.9%@10.
Automatic update strategy for real-time discovery of hidden customer intents in chatbot systems
2022, Knowledge-Based Systems
Citation Excerpt :
Textual data composed of personal, emotional, and economic aspects were generated through individual interviews. Amato et al. [14] proposed an intelligent recommendation system based on multi-agent systems. The aim was to develop a route planner for cultural visits, for example, in museums.
Model degradation is still a challenge in real-time applications such as chatbot systems. This work refers to a webchat service of a Brazilian energy utility company, whose central part is composed of a supervised model and a question-and-answer list. User queries not met by it go to an NLP-based clustering model, responsible for identifying unknown customer intents. Manual labeling is impractical in this case due to the large volume of data. This work proposes an automatic update strategy for this clustering model, necessary due to changes in customer behavior from time to time. A series of experiments showed considerable temporal variation in the number of user queries per customer intent, the allocation of queries from unknown intents to few clusters, and larger relative variations in cluster sizes for unknown rather than known intents over time. Based on these findings, a monitoring metric, together with a cut-off point, was proposed to be used as a trigger for updating the clustering model. This update task was demonstrated in a real situation, from which the discovery of new customer intents was confirmed by experts. This resulted in a significant recovery rate of 13.9% (2251 messages), as 85.1% of the queries are already answered promptly by the central part of the chatbot system. These findings are valuable for the company to improve service quality and, ultimately, customer satisfaction.
Personalized recommendation of collective points-of-interest with preference and context awareness
2022, Pattern Recognition Letters
Citation Excerpt :
Lyu et al. [17] proposed a personalized POI recommendation framework, called iMCRec, which recommends POIs by integrating user preferences on geographical, category, and attribute criteria with personalized weights. Amato et al. [18] introduced a methodology integrating recommendation facilities with agent-based planning techniques in order to implement a planner of routes within cultural sites as museums. Casillo et al. [19] developed a recommender system that suggests points of interest and related services according to both the profile of the tourist and contextual aspects.
With the popularity of mobile devices, location-based services, such as Foursquare and Facebook, have attracted increasing attention in recent years. Users share their interests anytime, anywhere with their friends on the social network. Point-of-Interest (POI) recommendation, as one of key services on Location-Based Social Networks (LBSNs), can effectively enhance users’ experience especially when they travel in a new city. Previous studies have made great success on POI recommendation by employing geographical influence and user preference. However, it is believed that the human decision on where to visit is very complex and involves the comprehensive factors such as POI popularity, POI location, user trajectory, and time context. In this paper, we propose a collective POIs recommendation framework which leverages the individual latent preference and contextual information. Firstly, to recommend top-K initial POIs, a scoring prediction model is constructed, which considers the influence of similarity, popularity and location of POIs. Furthermore, a next POI recommendation model based on personalized transfer probability is proposed, and the initial POI recommendation is combined to calculate the user’s score on the next POI. Extensive experiments based on real datasets collected from Foursquare demonstrate the proposed framework outperforms the state-of-art ones.
TCD-CF: Triple cross-domain collaborative filtering recommendation
2021, Pattern Recognition Letters
Citation Excerpt :
Yu et al. [18] propose a model (user-item-side) by inferring intrinsic user and item features, where they employ the bi-orthogonal tri-factorization model to infer the intrinsic user and item features from the user-side and item-side auxiliary domain respectively. Amato et al. [19] propose a methodology (user-item-side) integrating recommendation facilities with agent-based planning techniques in order to implement a planner of routes within cultural sites as museums. However, these approaches only use single-domain or dual-domain information to recommend target domains.
Recently, data sparsity is still one of the critical problems faced by recommendation systems. Although many existing methods based on cross-domain can alleviate it to a certain extent, these methods only use the information of single-domain (e.g., user-side, item-side and rating-side) or dual-domain (e.g., user-rating-side, user-item-side and item-rating-side) to make recommendations, which results in performance degradation. In this paper, we propose a triple cross-domain collaborative filtering method to alleviate data sparsity, named TCD-CF. In TCD-CF method, the triple-side intrinsic characteristics are first obtained by using the joint nonnegative matrix factorization to integrate the user-side, item-side and rating-side domain knowledge. Then the extended codebook (as knowledge to transfer) based on these intrinsic characteristics is constructed by using the orthogonal nonnegative matrix tri-factorization. Finally, the codebook-based transfer method for cross-system CF is applied into the source domain and target domain to predict the missing ratings and perform recommendation in the target domain. Extensive experiments on two real-world datasets demonstrate that the proposed method outperforms the state-of-the-art methods for the cross-domain recommendation task.
Pattern recognition and artificial intelligence techniques for cultural heritage
2020, Pattern Recognition Letters
This paper is the editorial of the virtual special issue (VSI) “Pattern recognition and artificial intelligence techniques for cultural heritage”, of which the authors of this paper have been the guest editors. It aims to bring together the work of experts from the fields of pattern recognition and artificial intelligence and that of cultural heritage. This multidisciplinary subject covers a wide spectrum spanning from the study of the cultural heritage to the development of tools based on PR/AI techniques for cultural heritage analysis, reconstruction and understanding. The papers included in this special issue allowed us to highlight the advances on this subject from a wide-angle perspective, as well as to stimulate new theoretical and applied researches for better characterizing the state of the art in this domain.

View all citing articles on Scopus

View full text

An agent-based approach for recommending cultural tours

Highlights

Abstract

Introduction

Section snippets

Recommender systems for cultural heritage

Recommendation

The recommendation strategy

Case studies

Conclusions and future works

CRediT authorship contribution statement

Declaration of Competing Interest

Artif. Intell.

Knowl.-Based Syst.

Future Gener. Comput. Syst.

Comput. Hum. Behav.

Artif. Intell.

A multimedia recommender integrating object features and user behavior

Multimed. Tools Appl.

A multimedia recommender system

ACM Trans. Internet Technol.

Snops: a smart environment for cultural heritage applications

Proceedings of the Twelfth International Workshop on Web Information and Data Management

Recommending multimedia visiting paths in cultural heritage applications

Multimed. Tools Appl.

Personalization and the web from a museum perspective

Museums and the Web