China
Africa
Explore chapters and articles related to this topic
Python API Modules for Machine Learning and Arduino
Published in Amartya Mukherjee, Nilanjan Dey, Smart Computing with Open Source Platforms, 2019
Amartya Mukherjee, Nilanjan Dey
Python library has a good range of API support for artificial intelligence and hardware interfacing. The machine learning application is one of the most popular packages that is widely implemented using python scikit learning tools. In this chapter, our main objective is to emphasize different useful packages that are related to machine learning, Arduino, Data Science, and other technologies. We are here trying to give the insights of those packages with some useful examples. One of the helpful entities in this context is PyPI, the Python Package Index, which is a software repository for Python programming.
Cloud Computing Development Tools
Published in Sunilkumar Manvi, Gopal K. Shyam, Cloud Computing, 2021
Sunilkumar Manvi, Gopal K. Shyam
The pip utility is used to manage package installation from the PyPI archive and is available in the python-pip package in most Linux distributions. Each OpenStack project has its own client, so depending on which services your site runs, we need to install some or all of the packages: python-novaclient (nova CLI), python-glanceclient (glance CLI), python-keystoneclient (keystone CLI), python-cinderclient (cinder CLI), python-swiftclient (swift CLI), and python-neutronclient (neutron CLI).
Automated Detection of Anti-National Textual Response to Terroristic Events on Online Media
Published in Cybernetics and Systems, 2022
Megha Chaudhary, Sachin Vashistha, Divya Bansal
In our dataset, in addition to Hindi vocabulary words, we have words that are back-transliterated from Roman to Hindi native script. These words stand out of vocabulary for the pre-trained mBERT model. We identify the most frequently used words in our dataset, which are added to the existing vocabulary of mBERT. The word embeddings of the model are updated according to the new words added, and then the model is fine-tuned with the enhanced vocabulary to learn embeddings of the new words. Thus, the pre-trained mBERT is extended with the vocabulary by adding words from our back-transliterated code-mix Hindi-English corpus. The process can be visualized in Figure 2. We used the hugging face transformers library (Wolf et al. 2020) to download pre-trained mBERT and fine-tune it with extended vocabulary. The words added to the vocabulary were extracted from the dataset by tokenizing the text using Python Spacy library (“Spacy - PyPI” n.d.).
Which way is the bookstore? A closer look at the judgments of relative directions task
Published in Spatial Cognition & Computation, 2019
Derek J. Huffman, Arne D. Ekstrom
We applied a metric transformation to each participant’s map using the package nudged (version 0.3.1; https://pypi.python.org/pypi/nudged) within Python (version 2.7.10). The metric transformation included translation, rotation, and uniform scaling. Uniform scaling preserves the angles between stores, which is motivated here by the fact that we used a square environment; furthermore, this method requires fewer assumptions than affine transformations that include nonuniform scaling, reflection, or shearing. To assess map-drawing performance, we calculated the bidimensional regression coefficient between each participant’s metric-transformed map and the actual map coordinates using the R package BiDimRegression (version 1.0.6; Carbon, 2013; also see: Tobler, 1994; Friedman & Bernd, 2003). To assess whether there was a relationship between performance on the map-drawing task and the JRD task, we calculated the correlation coefficient between performance on the map-drawing task (as assessed by Fisher’s r-to-z transformed bidimensional regression coefficients) and the JRD task (as assessed by median error on the final block of the task).
Optimization of economic production scheduling under sales and operations constraints in consumer goods industries
Published in International Journal of Management Science and Engineering Management, 2023
Pierre-André M. Fruytier, Michael C. Georgiadis
The previous mathematical model is a NLIP formulation, which can potentially be solved using exact NLIP algorithms. Such algorithms are typically characterized by high computational costs even for medium-sized problems. Moreover, they cannot always guarantee a global optimal solution. To overcome this complexity, an evolutionary algorithm is used for the optimization. This algorithm is implemented by the first author using the following open-source Python libraries (Python Package Index - PyPI., n.d..), i.e. Pandas, NumPy, and geneticalgorithm, successfully used for earlier consulting work. The decision support system relies on the standard code of geneticalgorithm so schedulers can later benefit from the most recent releases of the library.