Data Science And Multiple Criteria Decision Making Approaches In Finance Applications And Methods 1st Ed 2021 Gkhan Silahtarolu instant download
Data Science And Multiple Criteria Decision Making Approaches In Finance Applications And Methods 1st Ed 2021 Gkhan Silahtarolu instant download
https://ebookbell.com/product/data-science-and-multiple-criteria-
decision-making-approaches-in-finance-applications-and-
methods-1st-ed-2021-gkhan-silahtarolu-36529746
https://ebookbell.com/product/data-science-and-analytics-for-smes-
consulting-tools-practical-use-cases-1st-edition-afolabi-ibukun-
tolulope-46295584
Data Science And Analytics For Smes Consulting Tools Practical Use
Cases 1st Edition Afolabi Ibukun Tolulope
https://ebookbell.com/product/data-science-and-analytics-for-smes-
consulting-tools-practical-use-cases-1st-edition-afolabi-ibukun-
tolulope-46364434
Data Science And Analytics Strategy An Emergent Design Approach
Kailash Awati
https://ebookbell.com/product/data-science-and-analytics-strategy-an-
emergent-design-approach-kailash-awati-47691368
https://ebookbell.com/product/data-science-and-algorithms-in-systems-
proceedings-of-the-6th-computational-methods-in-systems-and-
software-2022-vol-2-radek-silhavy-48874572
https://ebookbell.com/product/data-science-and-predictive-analytics-
biomedical-and-health-applications-using-r-2nd-2nd-edition-ivo-d-
dinov-49151856
https://ebookbell.com/product/data-science-and-emerging-technologies-
proceedings-of-daset-2022-yap-bee-wah-49459136
https://ebookbell.com/product/data-science-and-applications-for-
modern-power-systems-le-xie-50584588
Multiple Criteria Decision Making
Gökhan Silahtaroğlu
Hasan Dinçer
Serhat Yüksel
Series Editor
Constantin Zopounidis, School of Production Engineering and Management,
Technical University of Crete, Chania, Greece
This book series focuses on the publication of monographs and edited volumes of
wide interest for researchers and practitioners interested in the theory of multicriteria
analysis and its applications in management and engineering. The book series
publishes novel works related to the foundations and the methodological aspects
of multicriteria analysis, its applications in different areas in management and
engineering, as well as its connections with other quantitative and analytic disci-
plines. In recent years, multicriteria analysis has been widely used for decision
making purposes by institutions and enterprises. Research is also very active in the
field, with numerous publications in a wide range of publication outlets and different
domains such as operations management, environmental and energy planning,
finance and economics, marketing, engineering, and healthcare.
Serhat Yüksel
Business and Management
Istanbul Medipol University
Istanbul, Turkey
© 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
This book aims to evaluate different financial issues to reach sustainable economic
development. In this context, assessments were made on 6 different important issues.
In this way, it is aimed to identify the most important issues related to financial
issues. In this process, both data science and fuzzy multi-criteria decision-making
methods were considered. In this context, decision trees, artificial neural networks,
text mining, and methods such as AHP, ANP, DEMATEL, MOORA, TOPSIS, and
VIKOR were used. The most important contribution of the study is the use of these
methods, which are frequently preferred in the literature, in the same book.
This book aims to integrate data science applications, such as web mining, text
mining, and machine learning, with different significant majors like business, health,
economics, finance, and engineering. Within this framework, different perspectives
can be taken into consideration in this study. For example, machine learning
approach can be used to analyze financial performance or big data methodology
can be considered to evaluate the efficiency of the stock exchanges. Therefore, it can
be said that this study offers a novelty by focusing on various significant majors at
the same time. As a result, it is believed that this study makes a significant
contribution to the literature.
In this book, detailed analyses are made on 6 different issues related to financial
issues. In this context, financially important issues such as profitability in the
banking sector, the factors affecting economic development, the role of the state-
ments of the politicians on the financial system, and the factors affecting the
exchange rate risk are examined. As a result of detailed analyses, development
suggestions were made for each topic. Thanks to these suggestions, it will be
possible to reach a more effective financial system and sustainable economic growth.
Therefore, this book is intended to make an important contribution to the literature.
v
Acknowledgments
The authors would like to acknowledge the help and patience of their families in this
book process. Without their support, this book would not have become a reality.
Second, the authors wish to acknowledge the valuable contributions of the
reviewers regarding the improvement of quality, coherence, and content presentation
of chapters.
In addition, the authors would also like to acknowledge the valuable help of
Mr. Serkan Eti for his significant support about machine learning system.
Gökhan Silahtaroğlu
Hasan Dinçer
Serhat Yüksel
vii
Introduction
In this book, analyses related to financial issues have been made. In this context,
assessments were made on 6 different important issues. In this way, it is aimed to
identify the most important issues related to financial issues. In this process, both
data science and fuzzy multi-criteria decision-making methods were considered. In
this context, decision trees, artificial neural networks, text mining, and methods such
as AHP, ANP, DEMATEL, MOORA, TOPSIS, and VIKOR were used. The most
important contribution of the study is the use of these methods, which are frequently
preferred in the literature, in the same book.
In the study, firstly, the factors causing crises in developing and developed
countries have been tried to be determined. In this framework, decision tree and
fuzzy DEMATEL approaches are taken into consideration. In addition, the second
chapter is related to identifying the influencing factors of economic growth for both
developing and developed economies. For this purpose, decision tree approach and
fuzzy TOPSIS methodology are considered at the same time.
On the other side, the third chapter aims to estimate the factors affecting the
profitability of the Turkish banking sector. For this purpose, 34 different variables
were firstly determined by literature review. In the first stage of the analysis, decision
trees method is applied to select the most important variables. After that, fuzzy ANP
approach is used to weight these variables. Similarly, the fourth chapter tries to
understand the role of the politicians on the macroeconomic situation of the coun-
tries. For this purpose, the tweets of Donald Trump are taken into consideration. Text
mining approach is used to evaluate these tweets and mostly used words are
identified. After that, these keywords are ranked with the help of fuzzy VIKOR
approach according to their impacts on macroeconomic performance.
In addition to them, the fifth chapter tries to understand the main influence of the
politicians’ disclosure on the stock exchange index. In this context, a machine
learning model is built in order to understand the hidden patterns behind the daily
changes (rises and falls) of Dow Jones index. Moreover, in the final chapter, it is
aimed to determine the factors affecting the exchange rate risk of companies. In this
context, firstly, articles in the ScienceDirect database that contain exchange rate risk
ix
x Introduction
in their titles, abstracts, and keywords are provided. Single, double, and triple words
were identified in 152 different studies that were published after 2018, which met the
relevant criteria. As a result of the analysis of these words, 4 different criteria that
could affect the exchange rate risk were determined. The relevant criteria were then
weighted with fuzzy AHP.
This book aims to integrate data science applications, such as web mining, text
mining, and machine learning, with different significant majors like business, health,
economics, finance, and engineering. Within this framework, different perspectives
can be taken into consideration in this study. For example, machine learning
approach can be used to analyze financial performance or big data methodology
can be considered to evaluate the efficiency of the stock exchanges. Therefore, it can
be said that this study offers a novelty by focusing on various significant majors at
the same time. As a result, it is believed that this study makes a significant
contribution to the literature.
Target audience and potential users of this book are defined below.
• Researchers
• Academicians
• Policy makers
• Government officials
• Upright students in the concerned fields
• Members of chambers of commerce and industry
• Top managers of the companies
Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3 Determining the Ways to Increase Economic Growth of Developing
and Developed Economies: An Application with Data Mining
and Fuzzy TOPSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.3 An Application on Developing and Developed Economies . . . 58
3.3.1 Data Set and Variables . . . . . . . . . . . . . . . . . . . . . . . . 58
3.3.2 Fuzzy TOPSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3.3 Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4 Profitability Prediction of Turkish Banking Industry:
A Comparative Analysis with Data Science and Fuzzy ANP . . . . . 77
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.3 An Application on Turkish Banking Industry . . . . . . . . . . . . . 80
4.3.1 Data Set and Variables . . . . . . . . . . . . . . . . . . . . . . . . 80
4.3.2 Fuzzy ANP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.3.3 Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Analysis Details of Decision Tree Approach . . . . . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5 The Influence of the Politicians on Macroeconomic Performance:
An Analysis of Donald Trump’s Tweets . . . . . . . . . . . . . . . . . . . . 109
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.3 An Application on Donald Trump’s Tweets . . . . . . . . . . . . . . 112
5.3.1 Fuzzy VIKOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.3.2 Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Contents xiii
financial crisis. He has more than 200 scientific articles, and lots of them are indexed
in SSCI, Scopus, and Econlit. Also, he is the editor of many books that were
published by Springer and IGI Global.
Chapter 1
Introduction to Data Science and Machine
Learning Algorithms
Abstract Data science has gained importance since available data and hardware
facilities have been ubiquitous. Algorithms to process a huge amount of data and
extract information were developed decades ago. However, due to the lack of high-
capacity computers, it was not possible to use them on real-life data and problems.
Today, from finance to medicine data science plays an important role to solve
problems. Suffice it to say, machine learning algorithms are the core of this new
phenomenon besides data itself. Artificial neural networks, deep learning, Support
Vector Machines, Decision Tree Learning Models, and related algorithms have been
used successfully and yielded very important results recently. On the other hand, text
data have also gained importance being the fuel of machine learning in data science.
Especially the emergence of social media and communication technology contrib-
uted to the popularity of texts in data science. In this chapter, concise introductions
have been given about the most popular and also successful machine learning
algorithms. This chapter will be helpful for those readers who do not have enough
information about machine learning and its algorithms.
Neural networks are a set of algorithms designed to recognize patterns, inspired from
the human brain and biological neural networks (Glorot & Bengio, 2010). They
interpret raw input by labelling or clustering them by a kind of machine perception.
Data patterns that they perceive are numerical data stored in vectors such as pictures,
sounds, text or time series, and these data need to be converted into a certain data
structure to make sense.
Artificial neural networks help us cluster and classify data in a meaningful way
that directly solves real-life problems or help us solve those problems. You can think
of them as a clustering and classification layer of the data you store and manage.
They help to group unlabelled data by similarities between sample inputs and
classify the data when there is a labelled data set to work on. Neural networks can
also extract features to feed other algorithms for clustering and classification; so, you
can think of deep neural networks as components of larger machine learning
applications that include algorithms for reinforcing learning, classification, and
regression.
What kind of problems does deep learning solve, and more importantly, can it
solve our problem? To know the answer, we need to ask some questions: What
outputs do I care about? Are these outputs the labels that can be applied to the data
available: For example, spam, clean user or fraud, unsatisfied customer or happy
customer. Are there any data to extract these labels? So, can I find data that has been
tagged before, or should I create a tagged dataset?
Deep learning maps inputs over outputs and finds correlations between them. It is
known as “a universal function derivative and solver” because it can learn to use an
unknown function f (x) ¼ y between any input x and any output y, assuming that
they are all related (e.g., correlated). In the learning process, when the neural
network is fed with f (x) ¼ 3x + 12 or f (x) ¼ 9x 0.1, it finds the correct form
of the equation by converting f to work for any x to reach y.
Deep learning is a term for “stacked neural networks,” i.e., networks of several
layers (Silahtaroğlu, 2008). Layers are made up of nodes that represent and act like
neurons. Node refers to the place where the calculation takes place, as a neuron
makes activations in the human brain, a node makes activations when it encounters
sufficient stimulus. A node combines input from data with a set of coefficients or
weights that increase or decrease the effect of this input, thus assigning importance
(value coefficient) to the task-related inputs that the algorithm is trying to learn.
Neural networks learn the answers of questions as follows:
Which input is most useful? How can I classify the data set without errors? How
can I calculate the final result? What is the best equation to solve the given problem
in order to achieve the task? To do all these, an activation function is used for each
node. According to the value of the activation function, it is calculated whether the
signal will pass through the network and how much information pass through that
layer (Silahtaroğlu, 2009). Figure 1.1 illustrates a node (artificial neuron) and its
activation function. As it is seen a neuron first sums all values it received from the
previous layers and then activates it.
1.2 Neural Network Elements 3
Basically, in a simple artificial neural network, x1, x2, x3, . . . xm inputs are given
and arbitrarily defined w weights compose an f (x) function. Inputs are simply the
dataset. Hidden layers are made up of many hidden layers each has got many nodes
or neurons. All hidden layers are interconnected via neurons. Each neuron has its
own SUM and ACTIVATION functions (Hambarde et al., 2020). Output layers are
the labels. A deep learning model tries to learn by creating a huge equation by the
help of the hidden layers in order to establish a rational among labels. The model
finds an equation to solve a nonlinear and sometimes chaotic correlation between
inputs (data) and outputs (labels) (Fig. 1.2).
4 1 Introduction to Data Science and Machine Learning Algorithms
Besides the number of hidden layers, nodes, and type of activation functions there
are some other parameters to consider when establishing a deep learning model.
Loss Function: It is the main function to be solved. It may be maximization
or minimization problem. It may also be a certain value
approximation. A typical neural network uses a gradient
descent algorithm to update weights. There are many others
though. Typical loss functions used are Cross Entropy,
Cosine Proximity, Mean Squared Error, Sum of Squared
Errors, Cross Entropy, Negative Log likelihood.
Optimizer: In order to keep loss function at a minimum level, the
weights of all neurons should be streamlined. When the
algorithm is searching for the best solution, it has to find
the right direction for a fast and efficient downward
movement. Most common optimizer algorithms are Line
Gradient Decent, Conjugate Gradient Decent, and
Stochastic Gradient Decent. They employ the following
updaters: Nesterov accelerated gradient (NAG),
ADAGRAD, ADADELTASGD, ADAM, MOMENTUM,
and RMSPROP.
Learning Rate: It is also known as step-size and its value ranges between
0 and 1. Smaller learning rate takes a long time to reach the
minimum loss function because moves will be with smaller
steps.
Weight Initialization: Under normal circumstances and in the original model of
neural networks, initial weights are taken randomly. Recent
studies showed that taking initial values which are closer to
the final weights to found through the model dramatically
reduces the learning time. XAVIER (Glorot & Bengio,
2010), SIGMOID, RELU, Softsign are some of the
algorithms (functions), which are used to choose initial
weights.
SVM. SVM performs this process in the form of linear, hyperbolic, cubic, etc. This is
also called KERNEL (Oreški & Oreški, 2014).
The line in the Fig. 1.4 separates the two classes. As it is clearly seen, one class is
on one side of the other class members are on the other side of the hyperplane.
However, in the real life, data are not separated so evenly as in the figure.
Please see Fig. 1.5 to image how data may be separated from one another in terms
of classes given. All representations are done in two-dimensional hyperplane, which
is the easiest one and can also be achieved by data visualization. SVM achieves it for
multidimensional hyperplanes (Bai et al., 2019).
When achieving the separations SVM exercises some conversions:
For conversion, another dimension is added, as we call the z-axis. Let us assume
the value of the points in the z plane, w ¼ x2 + y2. In this case, it can be arranged as
the distance from point z to all other points. Now if we draw on the z-axis, a clear
separation appears, and a line can be drawn (Fig. 1.6).
Another Random Scribd Document
with Unrelated Content
THE VENETIAN WELL, HOLMHURST.
SASSO.
“I spent a week at Bordighera. Such varied points for walks! villages like
Sasso, which are just bright bits of umber colour amongst the tender grey
olives; little painted towns amongst the orange-gardens, like Dolceacqua,
with its pointed bridge and blue river and great deserted palace of the
Dorias. George Macdonald, a most grand old patriarch to look upon, is king
of the place. He writes constantly, and never leaves the house, except to see
a neighbour in need of help or comfort. One after another of his delicate
daughters has faded away, but his sons seem strong and well, and there are
several adopted children in the house, half in and half out of the family, but
all calling Mrs. Macdonald ‘Mama.’ It is a very unusual household, but
ruled in a spirit of love which is most beautiful. I dined with them, the
dining-table placed across one end of the vast common sitting-room. On
Sunday evenings he gives a sort of Bible lecture, which all the sojourners in
Bordighera may attend.
AT BORDIGHERA. [498]
AT REBEKAH’S WELL, NEAR S. REMO.
“But it is in changed, spoilt Rome that I have spent the last two months.
All picturesqueness is now washed out of the place, so that people who
have any interest about them now usually give it only a glance and pass on.
It has been delightful for me, however, that Miss Hosmer is settled in this
hotel, and that we dine together daily at a little round table, where she is a
constant coruscation of wit and wisdom. All day she is shut up in her studio,
which is closed to all the world, but she cannot have a dull time, by the
stories she has to tell of the workmen and models who are her only
companions. Here are a few of them, only they sound nothing without her
twinkling eyes and capital manner of telling:—
GLEN AT S. REMO.
“‘Minicuccia was an excellent model, but very jealous. “Have you seen
Rosa? What fine arms she has!” I said to her one day. “I have seen
Rosaccia” she replied, “and I should have thought, Signorina, that a lady of
your taste would have known better than to admire her arms. What are they
in comparison with really fine arms—with mine, for instance?”
“‘One day Minicuccia was at a café, and some one admired the legs of
another model. Forthwith she gathered up her petticoats, and danced with
her legs perfectly bare all about the place. She was not a bad woman; on the
contrary, she was a very moral one, and there was never a word against her,
but she wanted to show what fine legs were. The police, however, heard of
that escapade, and she was put in prison for a month afterwards for such an
offence against the decenza pubblica. Poor Minicuccia!
“‘Then there was Nana, whom Lady Marian (Alford) painted so often,
and whom she was so fond of. She was a magnificent woman. Dear Lady
Marian used to say, “I would give anything to be able to come into a room
with the grace and dignity of Nana.” Her dignity was natural to her. Another
model once said to me, “I met that Nanaccia; she was walking down the Via
Sistina as if it all belonged to her.”
“‘There was a very nice boy-model I had, Fortunato he was called. He is
dead now—died of consumption, for he was always delicate. One day he
said to me, “Last Sunday, Signorina, I went to the garden of the Cappuccini,
and it is such a garden!—quite full of fruit, the most beautiful fruit. And the
Fathers are so kind; they said I might eat as much as ever I liked; only think
of that, Signorina!”—“Well, that was kind indeed; but what sort of fruit was
it?”—“O, cipolle and lettuge,[499] Signorina—most delicious fruit.”
“‘Marietta was another model who came to me, a large handsome
woman. One day I said to her, “Now, Marietta, I want you to look sad—
tutta dolorosa.”—“What! lagrime, Signorina?”—“No,” I said, “only look
sad; but if I wanted lagrime, could I have them too?”—“Sì, Signora: basta
pensare a quel calzolajo chi m’a fatto pagare sette lire in vece di cinque, et
piango subito.”[500]
“‘Marietta had a brother who managed her little business for her. I asked
her if it would not be very easy for him to misappropriate a scudo now and
then. “Facile sì,” she said, “essendo fratello.”
“‘Mariuccia lived to be old, and many is the dinner and paolo I have
given her; but when she was fifteen or so, she was the model for Mr.
Gibson’s ‘Psyche borne by the Zephyrs.’ She was always a wonderful
model: no one could act or stand as she did.
“‘Then there was that woman who had the drunken husband, who used
to beat her. One night he came in late and fell down dead drunk across the
bed. She took her needle and thread, and sewed him up in the sheets so that
he could not move, and then she took a stick, and beat him so that he died
of it: she was imprisoned for some years for that, though.
“‘I asked one of the workmen what he did when every one was away.
“Why, Signorina, I have the studio to clean out.”—“Well, I suppose that
takes you half-an-hour; and what do you do then?”—“Ma, Signorina, sto a
sedere.”—“And after your dinner, what do you do then?”—“Sto ancora a
sedere, Signorina.”—“Well, and in the evening?”—“Ma, Signorina,
continuo di stare a sedere.”
“‘My man Gigi came to me the other day and said, “I went to the Acqua
Acetosa[501] last Sunday, Signorina, and I liked the water so much, I drank
no less than twenty fiaschi of it.”—“Well,” I said, “Gigi, that was a good
deal; I’ll get twenty fiaschi of it, and put twenty scudi down by them, and
then, if you can drink them all off, you shall have the scudi.”—“Well,
Signorina, perhaps I did exaggerate a little: now I come to think it over,
perhaps it was ten fiaschi I drank.”—“Well, do it again before me, and you
shall have ten scudi.” “Now, Signorina, you know I like to be precise,
perhaps it was six fiaschi I drank.”—“Well, do it again and you shall have
six scudi.”—“Well, I suppose it really was two fiaschi.”—“Oh, I could
drink that myself!”’
“You may imagine how entertaining stories like these—traits from the
life around one—make our little dinners, and afterwards we often go into
the Storys’ apartment close by, where the easy intellectual pleasant talk and
fun are always reviving. Besides, it amuses Mrs. Story, who is most sadly
ailing now, though her cheerfulness is an example. She says she comforts
her sleepless nights by the old distich—
“Nothing can describe the charm of Mr. Story’s natural bubble of fun
and wit, or the merry twinkle which often comes into his eye, even now, at
moments when his wife’s illness does not make him too anxious.[502] He
and Miss Hosmer are capital together. It is difficult to say what are their
‘projecting peculiarities,’ as Dr. Chalmers would have called them, they
have so many; but they are all of a perfectly delightful kind.
“‘Well, what’s the news, Harriet?’ he said as we went in to-night. ‘Why,
that I am going to be married.’—‘What! to the Pope?’—‘Yes, only I didn’t
want it to get out till he announced it himself.’
“‘An American was looking at my statue of Canidia the other day,’ said
Mr. Story, and exclaimed—“Ah! Dante, I suppose, or is it—Savonarola?”
Another man who came to my studio said, “Mr. Story, have you baptized
your statue?”—“Why, yes,” I said; “generally we think of the name first,
and then we set to work in accordance with it.”—“Well,” he said, “there’s
some as doos, and there’s some as doosn’t.”’
“Mrs Story was very amusing about an Italian who wanted a portrait of
his father very much, and came to an artist she knew and asked him to paint
it. The artist asked, ‘But when can I see your father?’—‘Oh, you can’t see
him: he’s dead.’—‘But how can I paint him, then?’—‘Well, I can describe
him, and he was very like me: I think you can paint him very well.’ So the
artist painted away, according to the description, as well as he could. When
he had finished the portrait he sent for the son, anxious to see if he would
find any likeness. The son rushed up to the picture, knelt down by it, was
bathed in tears, and sobbed out, ‘O padre mio, quanto avete sofferto, o
quanto siete cambiato: O non l’aveva mai riconosciuto.’
“Mr. Story says that when Othello was performed at Rome, he saw it
with an Italian friend, who said afterwards, ‘Convengo che ci sono qualche
belle concette in questa dramma, ma fare tanto disturbo per un fazzoletto
non mi conviene.’
“Miss Hosmer told of a countryman who was asked what he thought of a
train, for he had just seen one for the first time—seen it as it was entering a
tunnel. ‘Well,’ he said, ‘it was just a black monster with a goggle eye, and
when it saw me, it gave a horrible shriek and ran into its hole.’
“I should like you to have heard Miss Hosmer’s recollections of Kestner,
whose name was so familiar to me in old Bunsen days. He died soon after
she first came to Rome, but she recollects him as always wearing his old red
studio cap. He knew he was dying, and when it was very near the end, he
said to those who were with him, ‘Now, my dear friends, it is a very sad
experience to see a person die: I must beg you to leave me: it is my great
wish to be alone, and you may come back in two hours.’ They came back in
two hours, and found him lying peacefully dead. That is a beautiful story, I
think. It was Kestner who, priding himself very much on his good English,
said to Lord Houghton, ‘Allow me to present to you my knee-pot (nipote).’
“Outside the charmed circle of Palazzo Barberini there is little now at
Rome but the most inferior American society. ‘We must stop at Milan, you
know, going back; there is a picture there by a man called Leonard Vinchey
we must be sure to see,’ said a neighbour at the hotel luncheon. And, ‘Mr.
Brown, sir, how’s Mrs. Brown’?—‘Well, she’s slim but round’ (meaning
weak but about): this is the sort of thing one hears.
“In this hotel is the intelligent Indian Princess Tanjore, with whom I
have spent several evenings very pleasantly. Her ‘lady’ is Miss Blyth, sister
of the Bishop of Jerusalem, and authoress of that capital novel ‘Antoinette.’
“Dear old Miss Garden, whom you will remember hearing of as the
kindest and most original of Scottish ladies, still lives at 64 Via Sistina.
‘How did you manage to boil the eggs so well, Maria, when you can’t tell
the clock?’ said Miss Garden to her old donna, ‘for the eggs are just
perfect.’—‘Why, I’ll tell you how it is,’ said Maria: ‘a lady I lived with
showed me how to do it. I just put them into the water, and then I say thirty-
three Credos, and then I know that they’re done.’
“With Miss Garden and Mrs. Ramsay I went one day to the curious little
early christian cemetery of S. Generosa, a lovely spot, where marble slabs
covering the graves of martyrs under Diocletian are still seen in a little
hollow surrounded by wild roses and fenochii.
“My room in this hotel looks out on the Barberini gardens, and the
splash of its fountain is an enjoyment. Its being lighted by electricity for the
King’s visit the other day was a type of the times, rather a contrast to twenty
years ago, when there were torches on every step of the great staircase to
welcome even a cardinal, and when not only the staircase, but the whole
street as far as S. Teresa, was hung with tapestries for the Prince’s funeral.
“On Ash-Wednesday I went, as I have always done here, to the ‘stations’
on the Aventine. It is still a thoroughly Roman scene. Before one reaches S.
Sabina, one is assailed by the chorus of old lady beggars seated in a double
avenue of armchairs leading up to the door, with ‘Datemi qualche cosa,
signore, per l’amore della Madonna, datemi qual’co;’ and behind them
kneel the old men—‘Poveri, poveretti cieci, signore,’ in brown gowns and
with arms stretched out alla maniera di S. Francesco. Spread with box is
the church itself, with its doors wide open to the cloistered porch and the
sacred orange-tree[504] seen in the sunny garden beyond. The Abbot is
standing there, and has his hand kissed by all the monks who arrive for the
stations, till a cardinal appears, after which he takes the lower place and is
quite deserted. Then we all hurry on to S. Alessio and its crypt, and then to
the Priorato garden, where, by old custom, we look through the keyhole of
the door, and see St. Peter’s down a beautiful avenue of bays.
“The passage of the Pope to the Sistine on his coronation anniversary
was a very fine sight. Borne along in his golden chair, with the white
peacock fans waving in front of him, and wearing his triple crown, Leo
XIII. looked dying, but gave his benediction with the most serene majesty,
sinking back between each effort upon his cushions, as if the end had
indeed come. Only his eyes lived, and lived only in his office; otherwise his
perfectly spiritualised countenance seemed utterly unconscious of the
thundering evvivas with which he was greeted, and which rose into a perfect
roar as he was carried into the Sala Regia. The potency of ‘Orders’ here is
so great, that my Swedish decoration not only gave me the best place, but I
took in two young men as my chaplain and equerry! After the Pope had
entered the Sistine, we sat in great comfort in the Sala Regia till he
returned, and then, as there was no one between us and the procession, we
saw all the individual faces of the old cardinals—how few of them the same
now as those I remember in the processions of Pius IX.
“There are no evvivas now for the comparatively young king with the
white hair and the ever-tragic countenance: the taxes are too great. I believe
that he can read, if no one else can, the handwriting on the wall which
foretells the doom of his southern kingdom. And yet personally no one
could be braver or more royal, and, where they detest the king, the people
honour the man. ‘Your king is at that house which has fallen down, helping
with his own hands to dig out that old man who is buried: he won’t leave till
the old man is safe,’ said Mrs. Story to her Italian maid Margherita. ‘Si,
Signora, casa di Savoia manca qualche volta di testa, mai di cuore;’ and it is
quite true. All one hears of the King’s self-abnegation is so fine. He used to
be quite devoted to smoking, but he was ill, and one day his physician told
him that it was extremely deleterious to him. He instantly took his cigar out
of his mouth, threw it into the back of the fire, and has never smoked again.
“The Pope’s secretary has just died of the influenza. Leo XIII. was much
attached to him, and is greatly distressed by his death. There is something
touching in the newspaper account of the Pope’s having refused to eat, and
his attendants having had to use qualche dolce violenze to make him do it.
“We have had two months of rain, only four fine days last week, in
which I went to the Crimera, to Fidenae, to Ostia, and to a touching and
beautiful Mass in the heart of the Catacomb of S. Praetextatus, where the
martyrs’ hymn was sung by a full choir upon their graves, its cadences
swelling through the subterranean church and dying away down the endless
rude passages, so long their refuge, and at last their place of death.
“And now I must stop. I am just come up from luncheon. ‘Wal, I guess
I’m stuffed, but I’m not appeased,’ said my neighbour as we came out; and
she was con rispetto parlando, as they say here—a lady.”
To Hugh Bryans.
“Rome, April 26, 1892.—How I wish you were here: how you would
enjoy it, though there is little to admire now in this much-changed Rome
beyond the extreme loveliness of the spring, with its Judas and May
flowers, and the golden broom of the Campagna. I have just been, with my
old friends Mrs. Ramsay and Miss Garden, to the Villa Doria to pick
anemones. There were thousands of them, and the ladies gathered them in
like a harvest. Their servant was told off to look after the violets. Their late
man, Francesco, said his was usually a very light place—‘ma nella
primavera, al tempo dei violette, e duro veramente.’
“I have seen little of the Easter ceremonies. On Holy Thursday I went to
St. Peter’s, and watched in the immense crowd for the extinction of the last
candle and beginning of the Miserere; but all the effect was lost and the
music inaudible from the incessant moving and talking. Afterwards there
was a fine scene at the blessing of the altar in the already dark church—the
procession, with lights, moving up and down the altar-steps, and then
kneeling all along the central aisle, whilst the relics were exhibited from the
brilliantly lighted gallery.
“Fifty-eight artisans and schoolmasters from the Toynbee Hall Institute,
with some of their wives, have been in Rome for the Easter holidays. On
Thursday I took them all over the Palatine, finding them most delightful
companions, and the most informed and interested audience I have ever
known. So since that I have been with them to the Appian Way, and Miss
Fleetwood Wilson kindly invited the whole party to tea at the old Palazzo
Mattei, unaltered through three hundred years. I made friends with many of
the party individually, and think that for really good, intelligent, high-
minded society, one should frequent the East End.
“What struck me most of all was the absence amongst them of the
scandal-talk which in our own society is so prevalent. ‘Consider how cheap
a kindness it is not to speak ill: it only requires silence,’ is an exhortation of
Bishop Tillotson. They remember this; we don’t.
“Do you recollect the pretty Miss Cators? With them and some pleasant
Americans, and Lanciani the famous archaeologist, I have been up Monte
Cavi. Lanciani was most delightful, and told us about everything in a way
which had all the enthusiasm and colour without the dry bones of
archaeology, and oh! what lilies, violets, cyclamen, narcissus, covered the
woods. Another day he lectured on old Fidenae, standing aloft on the
ancient citadel, with all his listeners in groups on the turf around him, and
afterwards they all had luncheon—still in scattered groups: it was like the
pictures of the miracle of the loaves and fishes.
To Miss Leycester.
“Cadenabbia, May 13.—At Florence, I went with the Duchess of
Sermoneta and Lady Shrewsbury to spend an evening with the grand old
family of Torrigiani, in the palace where the four sons, their wives, and
children innumerable, live with their charming mother, the Marchesa
Elisabetta, in perfect harmony and love; and another day went out to Poggio
Gherardo, a grand fortified villa, approached through half-a-mile of roses,
where the Ross’s now live. Then I was half a day at Padua, visiting it as a
tourist after many years, with my own book as a guide, and a most
delightful book I thought it!
“At Venice, I went to see ‘Pen Browning’ at the Palazzo Rezzonico, his
most beautiful old palace, full of memorials of Pope Clement XIII. The son
Browning has no likeness to either father or mother: he has worked hard,
both as painter and sculptor, and has a good portrait as well as a bust of his
father, from his own hands. There were many relics of his parents and their
friends, amongst them a sketch by Rossetti of Tennyson reading one of his
own poems to them, with an inscription by Mrs. Browning. ‘Pen’ was going
off to his house at Asolo, a place which his father first brought into notice
when he walked there and wrote ‘Pippa Passes.’
“Calling on a Mrs. Bronson in a neighbouring house, I met a young lady
with fluffy hair, a Countess Mocenigo. ‘My dear, how many Doges had you
in your family?’ said Mrs. B. ‘Seven,’ she answered, and there really were
seven Doges of the name Mocenigo, besides all those from whom she was
descended by the different marriages of her ancestors.
“Venice is still as full of odd stories as when my sister went to a party
there, and was surprised because the oddly dressed old lady by her side
never answered when she spoke, and then found she was made of wax.
Most of the company were, being ancestors present thus in the family life of
the present. Recently a lady named Berthold has lived at Venice who was of
marvellous beauty and charm. All the society flocked to her parties. One
evening she invited all her friends as usual. They found the palace
splendidly lighted, and listened to the most exquisite music. At the close of
the evening, curtains which concealed a platform at the end of the principal
room were drawn aside, and within, the beautiful hostess was seen, seated
on a throne, and sparkling with jewels, in all her resplendent loveliness.
And then, as she waved a farewell to all present, the curtains were suddenly
drawn, and she disappeared for ever. No human eye has seen her since. She
had observed signs, unperceived by others, that her beauty was beginning to
wane!
ebookbell.com