Professional Documents
Culture Documents
Download ebook New Perspectives On Enterprise Decision Making Applying Artificial Intelligence Techniques 1St Edition Julian Andres Zapata Cortes Editor Giner Alor Hernandez Editor Cuauhtemoc Sanchez Ramirez Editor online pdf all chapter docx epub
Download ebook New Perspectives On Enterprise Decision Making Applying Artificial Intelligence Techniques 1St Edition Julian Andres Zapata Cortes Editor Giner Alor Hernandez Editor Cuauhtemoc Sanchez Ramirez Editor online pdf all chapter docx epub
Download ebook New Perspectives On Enterprise Decision Making Applying Artificial Intelligence Techniques 1St Edition Julian Andres Zapata Cortes Editor Giner Alor Hernandez Editor Cuauhtemoc Sanchez Ramirez Editor online pdf all chapter docx epub
https://ebookmeta.com/product/chess-explained-
the-c3-sicilian-1st-edition-sam-collins/
https://ebookmeta.com/product/starting-out-the-c3-sicilian-1st-
edition-john-emms/
https://ebookmeta.com/product/the-paper-issue-83-1st-edition-
origamiusa/
The American Revolution 1774 83 2nd Edition Daniel
Marston
https://ebookmeta.com/product/the-american-
revolution-1774-83-2nd-edition-daniel-marston/
https://ebookmeta.com/product/bnf-83-british-national-formulary-
march-2022-joint-formulary-committee/
https://ebookmeta.com/product/english-grammar-exercises-with-
answers-part-4-your-quest-towards-c2-1st-edition-daniel-b-smith/
https://ebookmeta.com/product/patriot-vs-loyalist-american-
revolution-1775-83-combat-1st-edition-si-sheppard/
https://ebookmeta.com/product/british-light-infantryman-vs-
patriot-rifleman-american-revolution-1775-83-1st-edition-robbie-
macniven/
Studies in Computational Intelligence 966
New Perspectives
on Enterprise
Decision-Making
Applying Artificial
Intelligence
Techniques
Studies in Computational Intelligence
Volume 966
Series Editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
The series “Studies in Computational Intelligence” (SCI) publishes new develop-
ments and advances in the various areas of computational intelligence—quickly and
with a high quality. The intent is to cover the theory, applications, and design
methods of computational intelligence, as embedded in the fields of engineering,
computer science, physics and life sciences, as well as the methodologies behind
them. The series contains monographs, lecture notes and edited volumes in
computational intelligence spanning the areas of neural networks, connectionist
systems, genetic algorithms, evolutionary computation, artificial intelligence,
cellular automata, self-organizing systems, soft computing, fuzzy systems, and
hybrid intelligent systems. Of particular value to both the contributors and the
readership are the short publication timeframe and the world-wide distribution,
which enable both wide and rapid dissemination of research output.
Indexed by SCOPUS, DBLP, WTI Frankfurt eG, zbMATH, SCImago.
All books published in the series are submitted for consideration in Web of Science.
New Perspectives
on Enterprise
Decision-Making Applying
Artificial Intelligence
Techniques
Editors
Julian Andres Zapata-Cortes Giner Alor-Hernández
CEIPA Business School Division of Research and Postgraduate
Fundación Universitaria CEIPA Studies
Sabaneta, Colombia Instituto Tecnológico de Orizaba
Tecnológico Nacional de México
Cuauhtémoc Sánchez-Ramírez Orizaba, Mexico
Division of Research and Postgraduate
Studies Jorge Luis García-Alcaraz
Instituto Tecnológico de Orizaba Autonomous University of Ciudad Juarez
Tecnológico Nacional de México Ciudad Juárez, Chihuahua, Mexico
Orizaba, Mexico
Division of Research and Postgraduate
Studies
Tecnológico Nacional de México/Instituto
Tecnológico de Ciudad Juárez
Ciudad Juárez, Chihuahua, Mexico
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
With the increase in computing power produced in the last decade, the ability to
analyze large datasets and the development of new algorithms have produced the
possibility to develop mechanisms that allow decision-making processes that in past
years were thought impossible to be carried out. These mechanisms have the capacity
to surpass the human capacity for these tasks with the ability to carry out the anal-
ysis autonomously and to learn from themselves, allowing them to be conceived as
artificial intelligence.
Artificial Intelligence (AI) techniques cover the automation of cognitive and phys-
ical tasks. These techniques help people to perform tasks faster and make better deci-
sions. It enables the automation of decision-making process without human inter-
vention. AI techniques can enhance automation by reducing intensive human labor
and tedious tasks. There are more forms in which Artificial Intelligence is making
a difference for enterprises in decision-making marketing, customer relationship
management, recommender systems, problem-solving, opinion mining, augmented
analytics, to mention but a few.
In marketing, it is necessary to understand customer needs and desires and aligning
products to those needs and desires. A handle on changing customer behavior is vital
to make the best marketing decisions. AI simulation and modeling techniques provide
reliable insight into the consumers’ persona. This will help predicting consumers’
behavior. Through real-time data gathering, trend analysis, and forecasting, an AI
system can help businesses make insightful marketing decisions. Furthermore, orga-
nizations can identify a consumer’s lifetime value with the help of AI’s buyer persona
modeling. It can help organizations manage multiple inputs. During a complex
decision-making process, AI can efficiently manage and control different factors
at the same point in time. It can source and process large amounts of data within
minutes while providing valuable business-based insights. While we humans face
decision fatigue, algorithms do not have such limitations, which make AI-based
decisions faster and better. Also, AI techniques have provided businesses invalu-
able insight about consumers, which helps them enhance their communication with
the consumers. It also helps retailers predict product demand and respond to it
quickly. For this end, opinion mining helps businesses understand why people feel
v
vi Preface
the way they feel. Most often a single customer’s concerns might be common among
others. When sufficient opinions are gathered and analyzed correctly, the information
gleaned will help organizations gauge and predict the concerns of the silent majority.
AI has improved this mining process through automation, which is quicker and more
reliable, helping organizations in making critical business decisions.
In e-commerce, an AI system learns a consumer’s preference based on “explicit”
or “implicit” feedbacks. This kind of systems are called recommender systems. A
recommender system can provide information helping the organization to reduce
bounce rate and craft better customer-specific targeted content. Wise business deci-
sions are made when business executives and decision-makers have reliable data
and recommendations. AI not only improves the performance of both the individual
members of the team but also the competitive edge of the business.
In medicine, the artificial intelligence allows the improvement of many tasks,
ranging from better management of resources such as rooms and personnel, as well
as achieving more accurate and faster medical diagnoses, such as computer-assisted
radiologies or the use of robots to perform less invasive surgeries. In the service
industry, AI is used in automated online assistants that allow faster and lower cost
service, which has the ability to learn from past interactions, thus improving customer
service.
Based on this understanding, there are many other applications of artificial intelli-
gence in many knowledge fields, and this is the main reason and the core of the present
book, which is aimed to disseminate current trends among innovative and high-
quality research regarding the implementation of conceptual frameworks, strate-
gies, techniques, methodologies, informatics platforms, and models about Enter-
prise Decision-Making Applying Artificial Intelligence Techniques. The specific
objectives of the book can be summarized as follows:
• Generate a dissemination venue from both the academia and the industry in the
topics studied in the book, presenting cases of new approaches, applications,
methods, and techniques of the application of Artificial Intelligence in enterprise
decision-making.
• Generate a collection of theoretical and practical research works in the field of
Artificial Intelligence Techniques applied in enterprise decision-making.
• Establish the state of the art in the field of Artificial Intelligence Techniques
applied in enterprise decision-making.
This book is composed of a set of chapters, each of the kind of regular research
papers. These works have been edited according to the norms and guidelines of
Springer Verlag Editorial. Several calls for chapters were distributed among the
main mailing lists of the field for researchers to submit their works to this issue. 25
expressions of interest in the form of abstracts were received in total, which were
subject to a screening process to ensure their clarity, authenticity, and relevancy to
this book. These proposals came from several countries such as Colombia, Mexico,
Spain, Perú, and Ukraine.
After the abstract reviewing process, 25 proposals were accepted and asked to
submit full versions. These versions were reviewed by at least two pairs in order to
Preface vii
ensure the relevance and quality of the documents. After this process, 21 chapters
were finally accepted for their publication once the corrections requested by the pairs
and the editors were completed.
The book content is structured in three parts: (1) Industrial Applications, (2)
Decision-Making Systems for Industry, and (3) Artificial Intelligence Techniques.
The chapters in each of these parts are as follows.
Part I Industrial Applications: This part contains seven chapters.
Chapter 1, entitled “Merging Event Logs for Inter-organizational Process Mining,”
presents a methodology to merge the event logs of the different partners of a collab-
orative business process, in order to serve as input for the process mining algorithm.
On the one hand, the methodology consists of a method set for searching the corre-
lation between events of the log of different partners involved in the collaboration.
These methods are implemented at the trace level and the activity level. On the other
hand, the methodology consists of a set of methods and rules for discovering process
choreography. From the knowledge gained by the above methods, message-type
tasks are identified and marked in each event log, and then using a formal set of
rules, the message task sub-type (send or receive) is discovered. Finally, links using
message sequence flow connectors between message tasks identified as pair activi-
ties in event logs are automatically defined. The proposed approach is tested using
a real-life event log that confirms their effectiveness and efficiency in the automatic
specification of message flows of the process choreography discovered, allowing to
build a collaborative business process model.
Chapter 2, entitled “Towards Association Rule-Based Item Selection Strategy
in Computerized Adaptive Testing,” proposes the integration of Association Rule
Mining as an item selection criterion in a CAT system. Specifically, we present the
analysis of association rule mining algorithms such as Apriori, FPGrowth, Predic-
tiveApriori, and Tertius into three datasets obtained from the subject databases to
know the advantages and disadvantages of each algorithm and choose the most suit-
able one to employ in an association rule-based CAT system that is being developed as
a Ph.D. project. We compare the algorithms considering the number of rules discov-
ered, average support and confidence, lift, and velocity. According to the experiments,
Apriori found rules with greater confidence, support, lift, and in less time.
Chapter 3, entitled “Uncertainty Linguistic Summarizer to Evaluate the Perfor-
mance of Investment Funds,” proposes a methodology to implement the uncertain
linguistic summarizer posed in Liu’s uncertain logic to measure the performance of
investment funds in the Colombian capital market. The algorithm extracts a truth
value for a set of linguistic summaries, written as propositions in predicate logic,
where the terms for the quantifier, subject, and predicate are unsharp. The linguistic
summarizer proves to be autonomous, successful, efficient, and close to human
language. Furthermore, the implementation has a general scope and could become a
data mining tool under uncertainty. The propositions found characterize with plenty
of sense the investment funds data. Finally, a corollary that allows accelerating the
obtention of the summaries is presented.
viii Preface
estimates, loss estimates, as well as fraud detection and prevention, in the entities in
charge of managing an electric network, with the focus of doing so in smart grids.
Part II Decision-Making Systems for Industry: This part contains seven chapters.
Chapter 8, entitled “Measuring Violence Levels in Mexico Through Tweets,”
proposes a novel way to evaluate what people say and how they feel about violence
by analyzing Twitter data. To do this, we describe a methodology to create an indi-
cator to denote social perception. This methodology uses technologies like Big Data,
Twitter analytics, Web mining, and Semantic Web by manipulating software like
ELK (Elasticsearch for data storage, Logstash for collecting data, and Kibana for
data visualization); SPSS and R for statistical data analysis; and Atlas.ti, Ghephi,
and Wordle for semantic analysis. At the end of the chapter, we show our results
with a word cloud, a social graph, and the indicator of social perception of violence
in Mexico at the federal entity and metropolitan zone levels.
Chapter 9, entitled “Technology Transfer from a Tacit Knowledge Conserva-
tion Model into Explicit Knowledge in the Field of Data Envelopment Analysis,”
presents a preservation model of production engineers’ tacit knowledge of Data
Envelopment Analysis (DEA). Their expertise was explicitly coded into a computer
system, and the model was developed by applying techniques and procedures from
the fields of engineering and knowledge management. Technology transfer enables to
solve the problem of selecting criteria and interpreting results with DEA techniques,
when the efficiency of similar organizations is compared using an efficient frontier
derived from non-parametric approximations of such techniques. Misunderstanding
the techniques leads to misinterpretations of DEA results. This model was created
by applying Knowledge Engineering, which enables to preserve and extend specific
experiences and expertise in time by means of computer solutions. The model had an
efficient and positive impact on strategic self-learning processes for the community
interested in production engineering, knowledge transfer, and management.
Chapter 10, entitled “Performance Analysis of Decision Aid Mechanisms for
Hardware Bots Based on ELECTRE III and Compensatory Fuzzy Logic,” proposes
two novel cognitive preference models viable for hardware with few memory cells
and small processing capacity. These are the Outranking Relations (or OR) and
Compensatory Fuzzy Logic (or CFL) which are two techniques that lack a study of
their hardware performance as intelligence modelers, and despite their simple defini-
tion and generalization capacity, hardware agents almost never use them. The chapter
analyzes the feasibility of implementing Outranking Relations (OR) and Compen-
satory Fuzzy Logic (CFL) in hardware platforms with low resources, highlighting
that competitiveness of the proposed tools and the arising of new research lines.
Chapter 11, entitled “A Brief Review of Performance and Interpretability in Fuzzy
Inference Systems,” presents a scoping review related to fuzzy logic emphasizing
in Compensatory Fuzzy Logic (CFL), Archimedean Compensatory Fuzzy Logic
(ACFL), and inference systems. It presents the literature analysis on surveys and
general reviews to contrast the scoping review. The chapter also presents the research
analysis through a case study, a comparison of compensatory fuzzy logic with other
fuzzy logic structures and other related works.
x Preface
Chapter 12, entitled “Quality and Human Resources, Two JIT Critical Success
Factors,” presents a structural equation model associating JIT elements of quality
planning and quality management with human resources and economic performance
in the context of Mexican maquiladoras (cross-border assembly plants). Results
indicate that even though there is no direct relationship between quality planning and
management and economic benefits, these variables are indirectly related through
human resources. In conclusion, human resources are key to achieving financial
success of maquiladoras.
Chapter 13, entitled “Operational Risks Management in the Reverse Logistics of
Lead-Acid Batteries,” develops an operational risk identification and prioritization in
the reverse logistics of lead-acid batteries in Colombia using the questionnaires and
FQFD (fuzzy quality function deployment) approaches. We identified operational
risks in the reverse logistics of lead batteries and a probability-impact matrix to define
which of these risks should be prioritized with FQFD. In this way, we established
the priority of the risks considered. These prioritized risks help organizations related
to this activity to develop action plans to mitigate these risks. Once the most critical
risks are defined, actions to mitigate or eliminate them were proposed.
Chapter 14, entitled “Dynamic Evaluation of the Livestock Feed Supply Chain
from the Use of Ethanol Vinasses,” proposes a conceptual design of the vinasse-based
livestock feed supply chain using system dynamics to identify the key variables of
the chain and assess how vinasse can be efficiently used to produce animal feed
and other products such as ethanol. Both industrial and scientific efforts are made to
find alternative uses for vinasse in energy generation, soil fertilization, and livestock
feed production. However, this last alternative has not been sufficiently explored.
The results demonstrate that a continuous supply of molasses ensures a contin-
uous production of ethanol, which in turn guarantees constant vinasse availability to
produce livestock feed.
Part III Artificial Intelligence Techniques: This part contains seven chapters.
Chapter 15, entitled “Comparative Analysis of Decision Tree Algorithms for Data
Warehouse Fragmentation,” presents the analysis of different decision tree algo-
rithms to select the best one to implement the fragmentation method. Such analysis
was performed under version 3.9.4 of Weka considering four evaluation metrics
(precision, ROC area, recall, and F-measure) for different selected datasets using
the SSB (Star Schema Benchmark). Several experiments were carried out using two
attribute selection methods, Best First and Greedy Stepwise, the datasets were pre-
processed using the Class Conditional Probabilities filter, and the analysis of two
datasets (24 and 50 queries) with this filter was included to know the behavior of
the decision tree algorithms for each dataset. Once the analysis was concluded, we
can determine that for 24 queries dataset the best algorithm was RandomTree since
it won in two methods. On the other hand, in the dataset of 50 queries, the best deci-
sion tree algorithms were LMT and RandomForest because they obtained the best
performance for all methods tested. Finally, J48 was the selected algorithm when
neither an attribute selection method nor the class probabilities filter is used. But, if
Preface xi
only the latter is applied to the dataset, the best performance is given by the LMT
algorithm.
Chapter 16, entitled “Data Analytics in Financial Portfolio Recovery Manage-
ment,” presents the application of data analytics and machine learning techniques
to predict the behavior of the loan default in a non-financial entity. Five classifica-
tion algorithms (neural networks, decision trees, support vector machines, logistic
regression, and K neighbors) were run on a dataset of credit behavior data. Decision
trees have shown the best prediction performance to determine whether a loan will
be paid or become irrecoverable after running five predictive models.
Chapter 17, entitled “Task Thesaurus as a Tool for Modeling of User Information
Needs,” describes some examples of task thesaurus usage in intelligent applica-
tions for adaptation to user needs. Task thesaurus is an element of user model that
reflects dynamic aspects of user’s current work. Such thesaurus is based on domain
ontology and contains the subset of its concepts that are dealt with user task. These
can be generated automatically by analysis of task description or with the help of
semantic similarity estimations of ontological concepts. Task thesaurus represents
personalized user view on domain and depends on his/her abilities, experience, and
aims.
Chapter 18, entitled “NHC_MDynamics: High-Throughput Tools for Simulations
of Complex Fluids Using Nosé-Hoover Chains and Big Data Analytics,” is focused
on the implementation of the NVT algorithm of molecular dynamics based on the
Nosé-Hoover thermostat chains with the high-performance computing technology
such as Graphical Processing Units (GPUs) and Big Data analytics for the generation
of knowledge to help understand the functioning of thermodynamic properties in
simulated systems of Lennard-Jones fluids, as well as the study of the behavior of
proteins, such as diabetes, to refine their structures to formulate and integrate them
into the improvement of food in an improved diet for people with this condition.
Chapter 19, entitled “Determination of Competitive Management Perception
in Family Business Leaders Using Data Mining,” seeks to determine competitive
management perception of family business leaders, in order to establish working
assumptions in new research and propose improvement and consolidation initiatives
for these types of companies. This non-probabilistic, intentional study applied an
instrument with 10 dimensions and 94 variables to a sample of 133 family business
leaders from an intermediate city and a large city in Colombia. Data collection was
achieved using supervised machine learning algorithms in the Python programming
language, using techniques such as Cronbach’s Alpha Test, KMO, Levene, Bartlett,
Discriminant Analysis, and Decision Trees. The results allow us to identify four main
components in 19 variables: Management and technology, Quality Management,
Compensation, and Country competitiveness.
Chapter 20, entitled “A Genetic Algorithm for Solving the Inventory Routing
Problem with Time Windows,” presents a genetic algorithm that allows to simultane-
ously optimize inventory allocation and transport routes to supply a set of customers
for a specific time horizon. This model allows to obtain a minimum total cost as a
result of a better combination of the inventories at customers’ facilities and the trans-
portation required to supply them. The proposed model and the algorithm developed
xii Preface
for its solution allowed to obtain significant savings compared to the routing opti-
mization to supply all customers in each period using the vehicle routing problem
model with time windows, which allows to optimize customers’ inventory and mini-
mize transport costs in each period. However, when this solution is compared with
the total distribution cost throughout the time horizon, it generates higher costs than
the solution generated by the IRP with time windows presented in this work.
Finally, Chap. 21, entitled “Emotion Detection from Text in Learning Environ-
ments: A Review,” introduces a literature review of text-based emotion detection
in learning environments. We analyze the main APIs and tools available today for
emotion detection and discuss their key characteristics. Also, we introduce a case
study to detect the positive and negative polarities of two educational resources to
identify the accuracy of the results obtained from five selected APIs. Finally, we
discuss our conclusions and suggestions for future work.
Once a brief summary of chapters has been provided, the editors would like to
express their gratitude to the reviewers who kindly accepted to contribute in the
chapters’ evaluation at all stages of the editing process.
This book is part of the effort of several organizations that support research as the
CEIPA Business School, National Council of Science and Technology in Mexico
(CONACYT), PRODEP, Tecnológico Nacional de Mexico/IT Orizaba, and the
Autonomous University of Ciudad Juarez (UACJ).
Also, we, the editors, want to express our gratitude to other organizations that share
the same purpose but are not mentioned here due to the long list. We also appreciate
and recognize the trust, effort, and collaboration and patience that all authors and
co-workers endorse us as editors. Finally, we are thankful and appreciate the Springer
Publishing experts, especially to Thomas Ditzinger for his invaluable support at any
moment, for providing good tips for editing from their experience, patience, and
advice to materialize this book.
xiii
Contents
xv
xvi Contents
xxv
xxvi List of Figures
1.1 Introduction
In this section, we delineate the current literature from pre-processing and merging of
logs, and end-to-end flow analysis of related processes by correlation methods. The
technique proposed in Claes and Poels (2014) consists of an algorithm that searches
for links between the data of the different partners suggesting the rules to the user on
how to merge the data, and a method for configuring and executing the merge; as well
as their implementation in the process mining tool ProM (Van Dongen et al. 2005).
The algorithm discovers links between the two event logs to indicate which data in
both event logs are considered to belong to the same process instance. The merging
rules are formulated in terms of relations between attribute values at an event or
trace level in the two event logs. The rules are based on four relationship operators
and two logical operators (and/or). The approach is tested through seven scenarios
involving two artificially created and three real-life event logs, for demonstration
and evaluation of the method and algorithm. The test results show positive results
for effectiveness and efficiency.
In Raichelson et al. (2017), they describe an automated technique to merge event
logs supporting two granularity levels. They generate a merged log with a focus
on the case view and a merged log that reveals the end-to-end instance view. The
matching of cases is based on temporal relations and text similarity, both structured
1 Merging Event Logs for Inter-organizational Process Mining 7
The process mining area that is aimed to discover, verify, and improve real business
processes from events available in process-aware information systems (van der Aalst
2016), that is, aims at extracting process knowledge from event logs. An event log
consists of a set of traces and each trace is composed of a sequence of events produced
by the execution of one case. Each event captures relevant data about the execution
of a given activity in the business process. Therefore, through the application of
process mining, organizations can discover how the processes were executed, verify
if the defined business rules and practices were followed, as well as gain insights into
bottlenecks, resource utilization, and other performance-related aspects of processes
(Rovani et al. 2015).
The merge criterion for our proposal is based on using cosine similarity to derive a
measure of traces similarity and activities similarity between two event logs. Cosine
similarity analyzes the similarity by measuring the cosine of the angle between the
two vectors of an inner product space (Han and Kamber 2012). It is measured by
the cosine of the angle between two vectors and determines whether two vectors are
pointing in roughly the same direction. The cosine similarity measure (Manning and
Raghavan 2008) cosSim(j,q) can be computed as follows (Eq. 1.1):
t
(Wi j .Wiq )
Cos Sim( j, q) = i=1 t (1.1)
t
( i=1 Wi j )2 .( i=1 Wiq )2
t
where i=1 (Wi j .Wiq ) is the sum of the scalar product of the weights of the term i
of
the word vector j by the weight of the term i of the word vector q, this divided by
t t
( i=1 Wi j )2 .( i=1 Wiq )2 , which is the square root of the sum of the weights of
the term i of the word vector j squared by the sum of the weights of the term i of the
word vector q squared.
The proposed approach is based on the following definitions, which make it
possible to formalize the phases of the methodology, which enables the fusion of
event logs and the discovery of the correlation of the messages exchanged in a
collaboration.
Definition 1.1 (Event log) Given an event log L, which is compound of a set of
cases T, where T are all instances of the business process execution. Each case T i
is composed of a finite set of activities A, which are the set of tasks contained in the
business process. These tasks are described by a set of At attributes (for example,
activity name, execution date, the user who executed it, among others), which detail
the context of the execution of the business process activities.
Definition 1.2 (The bag-of-words of cases) Refers to a vector of words that represents
each case T i in the event L. This bag-of-words is generated from the unique values
of the At attributes of each Aj T i activity.
1 Merging Event Logs for Inter-organizational Process Mining 9
1.4 Methodology
The proposed methodology for merging event logs and discovering process choreog-
raphy in an inter-organizational collaboration environment is composed of processing
of event logs, identifying the correlation between events, and collaboration discovery
phases. Figure 1.2 illustrates the procedure of the proposed methodology, where the
event records of the business processes involved in inter-organizational collaboration
are used as input to phase 1, and as an output from phase 3, the collaborative business
process model discovered.
This method consists of building bags-of-words per event log, which allows each
bags-of-words to be represented in a matrix. This procedure is performed by every
organization (one event log per organization) involved in the collaborative business
10 J. D. Hernandez-Resendiz et al.
process. For each event log L and L’ a matrix of bags-of-words BW, and BW ’ is
generated, where one row (a trace in the event log) of the matrix is a bag-of-word
of T L and T ’ L’, respectively. The length of each bag-of-words is equal to the
number of attributes that describe the activities contained in a trace. The matrices
BW and BW ’ are constructed as follows:
• Let us consider any case T i L.
• A bag-of-word BW i is built, with all the unique values of the At k attributes of
each task Aj T i , based on Definition 1.2.
• In BW i , stop-words are removed.
• The remaining words of BW i compose the bag-of-words T i .
• This procedure is carried out independently for the event logs L and L’.
In this method for each of the BW i bags-of-words that belong to the BW bag-of-
words, they are compared to all the bags-of-words that compose the BW ’. This is
performed using the cosine similarity according to Definition 1.4. The above allows
us to generate a scoring matrix with the similarity between the vectors of BW and
BW ’.
1 Merging Event Logs for Inter-organizational Process Mining 11
The selection of the even cases consists of finding the cases T L and T ’ L’ with
the closer the cosine value to 1. For each T i L case, the case T i ’ L’ with the
greater cosine similarity in the scoring matrix is selected. This allows us to identify
that in cases T L and T ’ L’ there is a relationship at the case level, according to
Definition 1.4. This process is performed for each trace of the event log L.
From the cases selected by Method 3, the tasks or events (activities) that contain
any coincident data are identified, allowing a relationship between the activities of
the traces to be defined, calculating their cosine similarity, that is, similarity at the
activity level, for which the following procedure is implemented:
• The attribute or attributes that can provide information on the relationship between
organizations are defined, for example, resource, timestamps, activity name, input,
or output objects.
• For each activity Aj T i and Aj ’ Ti’ of each pair of cases T i L and T i ’
L’ (which were selected in the previous method because their cosine similarity
value exceeds the U t threshold), two vectors of activities A and A’ are obtained,
which represent the bags-of-words BWA and BWA’ according to Definition 1.3,
respectively.
• For each of the BWAi bags-of-words, its measure of the cosine similarity distance
with all the BWA’ bag-of-words is calculated. This measurement allows us to
build a scoring matrix of the relationship identified at the activity level. If the
measure of similarity of a pair of the bags-of-words BWAi and BWA’i exceeds the
threshold U a , these are considered as message-type tasks. These tasks can be of
sub-type send or receive (BPMN 2.0 2011). Then, a vector that contains the pair
of message-type activities (PMA) that exceeded the threshold is generated.
SCREEN-DOOR
EYE FOR FISH- SPRING REPAIR HOLDS BUGGY
HOOK SINKER CURTAIN OVER
SPRING HINGE FOR TORN BUTTON-
BOX HOLE
DRY-BATTERY CONNECTOR
These Eight Examples of How an Ordinary Safety Pin can be Used for
Practical Purposes Are Suggestive of Many Others
It is surprising to note how many uses can be made of an ordinary
safety pin, both as an emergency-repair device, and for other
purposes. I found 40 simple uses for safety pins, a few of which are
shown in the illustration. Some of the many other uses are as a
chain, a candle holder, a spring, a cover holder for a kettle, and as a
pencil clip. Most of these can be made by twisting the safety pin with
a small plier. By using a cutting plier, many more useful kinks can be
devised.—George G. McVicker, North Bend, Neb.
Moving Heavy Objects with a Broom
The Broom Is Slippery and Rides over the Floor or Even Carpets without
Trouble
The finished billfold is shown with the cardcase side up; folded,
and with the bill compartment open, in Fig. 2. Begin the making of it
by cutting a pattern of stiff cardboard for the main portion, or body
part No. 1, as shown in Fig. 1.
Make a second pattern for body part No. 2, as shown in Fig. 1.
Trim the leather to fit the patterns, a straight edge of metal being
placed over the pattern as a guide for the knife. A close-grained
hardwood board, or a piece of sheet zinc, is suitable as a base for
cutting the leather.
Cut pieces of lining slightly larger than the patterns. If skiver is
used, it must be glued to the leather with a good quality of leather
glue, which should be pliable so as not to crack the leather. Do not
stretch the lining, but merely smooth it gently, as otherwise it may
cause the leather to buckle. Place the glued parts under a light
weight to dry. In fitting the silk lining into place, apply a thin stripe of
glue around the edges only, and permit them to extend beyond the
leather. Then trim them off neatly.
Part No. 1 is to be folded to form the back of the billfold and also
the two pockets for cards. The parts are glued as shown in Fig. 3.
Weight the billfold and permit the glue to dry. The sewing may then
be undertaken.
Fig. 2. The Photographs Show the Cardcase Side, the Folded Billfold, and
the Silk-Lined Money Compartment
Fig. 3. Glue the Parts Together Carefully along the Edges, and Fold the
Ends as Indicated
The neatness of the billfold will depend largely on the care with
which the sewing is done, and the finish on the edges. Draw a
straight line with a blunt-point tool along the edges to be sewed, ³⁄₁₆
in. from them. Mark the spacing for an inch of the stitches on a strip
of paper and transfer them to the stitching line. Punch the holes,
making certain that each is made with the awl held vertical, as
shown at D. Back the open ends with strips of cardboard when
punching holes in them. Heavy, waxed silk of a color to match the
leather is used for the sewing. The work may be done by hand
without a holder, clamped between two boards. Thread the two
needles and start them from the right side, as at E. Pass the needle
from the first hole through the second as at F. With the needles then
in position, as at G, continue this stitch. Tie the ends of the thread
neatly at the end of the stitching. Trim the edges and tool them to a
smooth finish with a suitable metal tool, or rod, heated moderately,
applying water with the finger tip.
¶The taper of a key is generally ¹⁄₈ in. per foot of length, and keys
having a head are preferable because of greater ease in removing
them. The width of a key is usually ¹⁄₄ the diameter of the shaft, plus
¹⁄₈ inch.
Lawn Mower Sharpened Efficiently with Simple
Rigging
With this device, one can quickly sharpen the lawn mower, or
perhaps earn money sharpening the neighbors’ machines. Because
several knives in a lawn mower work against the cutting blade, it is
difficult to sharpen a mower satisfactorily with a file or stone. A usual
method is to reverse the cutting-wheel cogs, turn the mower upside
down, and run it backward over the ground. By the use of grinding
paste on the cutting blade the cutting wheel tends to sharpen itself.
This process can be improved by supporting the mower, as shown at
A, the blocks permitting the mower’s driving wheels to rest outside
them. A block, B, supports the wooden roller. Remove the driving
wheels, and the gears which turn the cutting wheel. Take the gear
from one end and put it in the other end, replacing the driving wheel.
By revolving the latter backward, the cutting wheel will also turn
backward. Apply a paste of emery powder and lubricating oil to the
cutting blade, and adjust the blade so that it comes fairly into contact
with the cutting wheel. Turn the driving wheel backward, and
gradually tighten the adjustment until a good edge on the cutting
blade and the knives is produced.
A handle fixed on the driving wheel, as shown at C, makes turning
it easier. After sharpening, replace the gears, clean the bearings
thoroughly, and replace the driving wheels.—Ed. M. Hawes, Everett,
Wash.
Rubber Pads for Opening Screw Watch Bezel
The modern screw case for watches is primarily for the purpose of
keeping dust from getting into the delicate works. Unless one is
possessed of a strong grip, it is hard not only to set the case tight
enough to keep out dust, but also to open the watch when
adjustment is necessary. To make this work easy, a jeweler has, as a
part of his tool kit, several rubber dies to fit the different sizes of
watch cases. He made them from old rubber heels, cut to fit the
watch cases closely. Danger of breaking the crystal is also
overcome.—C. E. Drayer, Cleveland, Ohio.
Spring Roost Releases Poultry-House Door Latch
The Door is Opened Automatically When the Fowl Leave Their Roost