Trending March 2024 # Partial Auc Scores: A Better Metric For Binary Classification # Suggested April 2024 # Top 7 Popular

You are reading the article Partial Auc Scores: A Better Metric For Binary Classification updated in March 2024 on the website We hope that the information we have shared is helpful to you. If you find the content interesting and meaningful, please share it with your friends and continue to follow and support us for the latest updates. Suggested April 2024 Partial Auc Scores: A Better Metric For Binary Classification


Partial AUC (Area Under the Curve) scores are a valuable tool for evaluating the performance of binary classification models, particularly when the class distribution is highly imbalanced. Unlike traditional AUC scores, partial AUC scores concentrate on a specific region of the ROC (Receiver Operating Characteristic) curve, offering a more detailed evaluation of the model’s performance.

This blog post will dive into what partial AUC scores are, how they are calculated, and why they are essential for evaluating imbalanced datasets. We will also include relevant examples and a code example using Python to help make these concepts clearer.

Learning Objectives:

Learning the basics of AUC Scores.

How to calculate partial scores and their objective.

This article was published as a part of the Data Science Blogathon.

Table of Contents What do AUC Scores Mean?

AUC (Area Under the Curve) scores are a commonly used metric for evaluating the performance of binary classification models. The traditional AUC score calculates the area under the ROC (Receiver Operating Characteristic) curve, which plots the True Positive Rate (TPR) against the False Positive Rate (FPR) for all possible threshold values. The score ranges from 0.5 for a random model to 1 for a perfect model, with values closer to 1 indicating better performance.

The Drawback of AUC Scores

However, in real-world applications, the class distribution of the target variable can be highly imbalanced, meaning that one class is much more prevalent than the other. In these cases, the traditional AUC score may not provide a good evaluation of the model’s performance as it aggregates the performance overall threshold values and does not account for the imbalance in the class distribution.

Overcoming the Drawback

This is where partial AUC scores come into play. Unlike traditional AUC scores, they focus on a specific region of the ROC curve, providing a more granular evaluation of the model’s performance. This allows for a more accurate evaluation of the model’s performance, especially in cases where the class distribution is highly imbalanced.

For example, in a fraud detection problem, the partial AUC score can be calculated for the region where the FPR is less than a specific value, such as 0.05. This provides an evaluation of the model’s performance at catching fraud instances while ignoring the performance on the majority class instances. This information can be used to make informed decisions about which models to use, how to improve models, and how to adjust the threshold values for predictions.

Calculating Partial AUC Scores

Calculating partial AUC scores involves dividing the ROC curve into intervals and then computing the AUC for each interval. The intervals can be defined in terms of the FPR or TPR, and the size of the intervals can be adjusted to control the granularity of the evaluation. The partial AUC score for a specific interval is calculated as the sum of the areas of the rectangles formed by the interval boundaries and the ROC curve within that interval.

For example, to calculate the partial AUC score for the region where the FPR is less than 0.05,

we would first divide the ROC curve into intervals for the specific interval where the FPR is less than 0.05.

Then, we would calculate the sum of the areas of the rectangles formed by the interval boundaries and the ROC curve within that interval. This would give us the partial AUC score for that region.

In addition to the fraud detection example, partial AUC scores can be used in a variety of other real-world applications such as medical diagnosis, credit scoring, and marketing.

In medical diagnosis, they can be used to evaluate the performance of a model in detecting a specific disease while ignoring its performance on healthy individuals.

In credit scoring, they can be used to evaluate the performance of a model in detecting high-risk loan applicants while ignoring its performance on low-risk loan applicants.

In marketing, they can be used to evaluate the performance of a model in predicting which customers are most likely to purchase while ignoring its performance on customers who are unlikely to make a purchase.

In conclusion, partial AUC scores are an important tool for evaluating the performance of binary classification models, especially in cases where the class distribution is highly imbalanced. By focusing on a specific region of the ROC curve, partial AUC scores provide a more granular evaluation of the model’s performance, which can be used to make informed decisions about model selection, improvement, and threshold adjustment. Understanding them and how to use them is an important part of the evaluation process for binary classification models and can lead to more accurate and effective decision-making in various real-world applications.

It’s important to note that partial AUC scores are not a replacement for traditional AUC scores but rather a complementary tool to be used in conjunction with traditional AUC scores. While they provide a more nuanced evaluation of the model’s performance in specific regions of the ROC curve, traditional AUC scores provide a more holistic evaluation of the model’s overall performance.

When evaluating binary classification models, it’s best to use both traditional AUC scores and partial AUC scores to get a complete picture of the model’s performance. This can be done by plotting the ROC curve and calculating both the traditional AUC score and the partial AUC scores for specific regions of the curve.

Partial AUC Scores in Python

Now, let’s see how to calculate partial AUC scores in Python. The easiest way to calculate partial AUC scores in Python is by using the “roc_auc_score” function from the scikit-learn library. This function calculates the traditional AUC score by default, but it can also be used to calculate partial AUC scores by passing in the “curve” parameter.

For example, let’s say we have a binary classification model and its predictions on the test data. We can use the following code to calculate the traditional AUC score:

from sklearn.metrics import roc_auc_score y_true = [0, 0, 1, 1] y_scores = [0.1, 0.4, 0.35, 0.8] auc = roc_auc_score(y_true, y_scores) print('AUC:', auc)

To calculate the partial AUC score for the region where the FPR is less than 0.05, we can pass in the “max_fpr” parameter as follows

from sklearn.metrics import roc_auc_score y_true = [0, 0, 1, 1] y_scores = [0.1, 0.4, 0.35, 0.8] auc = roc_auc_score(y_true, y_scores, max_fpr=0.05) print('Partial AUC:', auc) Conclusion

In summary, it provides a more granular evaluation of the performance of binary classification models, especially in cases where the class distribution is highly imbalanced. Understanding and using them can greatly enhance the evaluation of binary classification models in imbalanced datasets.

By combining traditional AUC and partial AUC scores, we can get a complete picture of the model’s performance and make informed decisions about model selection, improvement, and threshold adjustment.

With the help of Python’s scikit-learn library, calculating partial AUC scores is a straightforward and convenient process.

When evaluating binary classification models, it’s best to use both traditional AUC scores and partial AUC scores to get a complete picture of the model’s performance.

The media shown in this article is not owned by Analytics Vidhya and is used at the Author’s discretion.


You're reading Partial Auc Scores: A Better Metric For Binary Classification

What Is Lstm For Text Classification?

easier. As you start exploring the field of deep learning, you are definitely going to come across words like Neural networks, recurrent neural networks, LSTM, GRU, etc. This article explains LSTM Python and its use in Text Classification. So what is LSTM? And how can it be used?

This article was published as a part of the Data Science Blogathon

What is LSTM?

LSTM stands for Long-Short Term Memory. LSTM is a type of recurrent neural network but is better than traditional recurrent neural networks in terms of memory. Having a good hold over memorizing certain patterns LSTMs perform fairly better. As with every other NN, LSTM can have multiple hidden layers and as it passes through every layer, the relevant information is kept and all the irrelevant information gets discarded in every single cell. How does it do the keeping and discarding you ask?


Tradition neural networks suffer from short-term memory. Also, a big drawback is the vanishing gradient problem. ( While backpropagation the gradient becomes so small that it tends to 0 and such a neuron is of no use in further processing.) LSTMs efficiently improves performance by memorizing the relevant information that is important and finds the pattern.

How Does LSTM Work in Python?

LSTM has 3 main gates:




Let’s have a quick look at them one by one.

1. FORGET Gate

This gate is responsible for deciding which information is kept for calculating the cell state and which is not relevant and can be discarded. The ht-1 is the information from the previous hidden state (previous cell) and xt is the information from the current cell. These are the 2 inputs given to the Forget gate. They are passed through a sigmoid function and the ones tending towards 0 are discarded, and others are passed further to calculate the cell state.

2. INPUT Gate

Input Gate updates the cell state and decides which information is important and which is not. As forget gate helps to discard the information, the input gate helps to find out important information and store certain data in the memory that relevant. ht-1 and xt are the inputs that are both passed through sigmoid and tanh functions respectively. tanh function regulates the network and reduces bias.

3. Cell State

All the information gained is then used to calculate the new cell state. The cell state is first multiplied with the output of the forget gate. This has a possibility of dropping values in the cell state if it gets multiplied by values near 0. Then a pointwise addition with the output from the input gate updates the cell state to new values that the neural network finds relevant.

4. OUTPUT Gate

The last gate which is the Output gate decides what the next hidden state should be. ht-1 and xt are passed to a sigmoid function. Then the newly modified cell state is passed through the tanh function and is multiplied with the sigmoid output to decide what information the hidden state should carry.

LSTM Python for Text Classification

There are many classic classification algorithms like Decision trees, RFR, SVM, that can fairly do a good job, then why to use LSTM for classification?

One good reason to use LSTM is that it is effective in memorizing important information.

If we look and other non-neural network classification techniques they are trained on multiple word as separate inputs that are just word having no actual meaning as a sentence, and while predicting the class it will give the output according to statistics and not according to meaning. That means, every single word is classified into one of the categories.

This is not the same in LSTM. In LSTM we can use a multiple word string to find out the class to which it belongs. This is very helpful while working with Natural language processing. If we use appropriate layers of embedding and encoding in LSTM, the model will be able to find out the actual meaning in input string and will give the most accurate output class. The following code will elaborate the idea on how text classification is done using LSTM.

Model Defining

Defining the LSTM python model to train the data on.

Code #model embedding_vector_features=45 model=Sequential() model.add(Embedding(voc_size,embedding_vector_features,input_length=sent_length)) model.add(LSTM(128,input_shape=(embedded_docs.shape),activation='relu',return_sequences=True)) model.add(Dropout(0.2)) model.add(LSTM(128,activation='relu')) model.add(Dropout(0.2)) # for units in [128,128,64,32]: # model.add(Dense(units,activation='relu')) # model.add(Dropout(0.2)) model.add(Dense(32,activation='relu')) model.add(Dropout(0.2)) model.add(Dense(4,activation='softmax')) print(model.summary()) Summary

Model: “sequential_2”

We define a sequential model and add various layers to it.


The first layer is Embedding layer. It representing words using a dense vector representation. The position of a word within the vector space is based on the words that surround the word when it is used. For eg. “king” is placed near “man” and “queen” is placed near “woman”. The vocabulary size is provided.

The next layer is an LSTM layer with 128 neurons. “embedded_docs” is the input list of sentences which is one hot encoded and every sentence is made of the same length. The activation function is rectified linear, which widely used. Any other relevant activation function can be used. “return_sequences=True” this is an important parameter while using multiple LSTM layer as it enables the output of the previous LSTM layer to be used as an input to the next LSTM layer. If it is not set to true, the next LSTM layer will not get the input.

A dropout layer is used for regulating the network and keeping it as away as possible from any bias. Another LSTM layer with 128 cells followed by some dense layers.

The final Dense layer is the output layer which has 4 cells representing the 4 different categories in this case. The number can be changed according to the number of categories.

Compiling the model using adam optimizer and sparse_categorical_crossentropy. Adam optimizer is the current best optimizer for handling sparse gradients and noisy problems. The sparse_categorical_crossentropy is mostly used when the classes are mutually exclusive, ie, when each sample belongs to exactly one class. The Summary explains all the details of the model.

Training the Model for LSTM Python

Now that the model is ready, the data to be trained it split into train data and test data. Here the split is 90-10. X_final and y_final are the independent and dependent datasets.


from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X_final, y_final, test_size=0.1, random_state=42,stratify=y_final)

The next step is to train the LSTM model using the train data, and the test data is used for validating. is used for this purpose.


The epochs are the number of times the process will be repeated. It can be 20,120, 2000, 20000, any number. The batch size is 64, ie, for every epoch, a batch of 64 inputs will be used to train the model. It mostly depends on how large the dataset is.


After training is completed, it’s time to find out the result and predict using the model.

1. Accuracy Code: results = model.evaluate(X_test,y_test)

The model is evaluated and the accuracy of how well the model classifies the data is calculated. “results” will have the accuracy score and the loss. In some cases increasing the number of epochs can increase the accuracy as the model gets trained better.

To use the trained model for predicting, the predict() function is used.

2. Predict Code:

The “embedded_docs_pred” is the list is words or sentences that is to be classified and is one-hot encoded and padded to make them of equal length. 

y_pred = model.predict(embedded_docs_pred) y_pred:

The output will have a list for every input (can be a word or a sentence). Every list has the predicted score for 4 classes. The maximum of them is the class for that respective input. As discussed above LSTM facilitated us to give a sentence as an input for prediction rather than just one word, which is much more convenient in NLP and makes it more efficient.

End Note

In conclusion, LSTM (Long Short-Term Memory) models have proven to be a powerful tool for text classification in Python. With their ability to capture long-term dependencies and handle sequential data, LSTM models offer improved accuracy in classifying text. By implementing LSTM models in Python, researchers and practitioners can leverage the strengths of this architecture to achieve better results in various text classification tasks, opening up new possibilities for natural language processing applications. Sign-up for our Blackbelt program if you want to learn more about it!

Frequently Asked Questions


Bitcoin Or Blockchain: Which Is A Better Choice For Business?

Blockchain and Bitcoin are distinctly apart, yet are very close to each other

Blockchain technology has gained huge momentum in the marketplace. For tech and IT businesses, integrating


has become an essential part of automating business processes seamlessly. As more and more businesses are adopting this cutting-edge technology, enterprise leaders are renovating traditional business strategies to gain an edge in the competitive market and keep up with the rising customer demands. Accompanying this technology comes cryptocurrency. As it is needless to mention,


are digital currencies that have revolutionized the very base of traditional finance and economy. The popularity of cryptocurrencies has experienced a substantial boost in recent times. With the emergence of this decentralized landscape, due to blockchain technology, major cryptocurrencies like Bitcoin have gained massive popularity among investors. But there are still several controversies revolving around this industry. One of the most prominent debates is about the differences and similarities between


and blockchain. 

Several investors consider blockchain and Bitcoin to be the same things. But they are quite distinct in nature but are also somehow closely related. When Bitcoin was launched as an open-source code, blockchain was wrapped up together and was started being addressed as the same solution. But since Bitcoin is the first-ever project on the blockchain, individuals generally mix up both of these together, and that’s how the entire confusion started. Bitcoin and blockchain are being exponentially used in businesses for a variety of use cases, but both have very distinct functionalities which might prove helpful for respective business leaders, based on their individual enterprise requirements. So to understand which is better for business operations, let’s dive in and explore the basic differences between blockchain and Bitcoin to clear out any confusion.

The Difference between Blockchain and Bitcoin

First, let’s start by focusing on the contextual differences between the two. Blockchain is the distributed ledger technology that records transactions between two parties with better efficiency. Whereas, Bitcoin is the world’s first and the largest cryptocurrency. Presently, there are several other major cryptocurrencies in circulation that are competing with and against each other to dethrone Bitcoin as the best


in the market. 

Bitcoin transactions are stored and transferred using a distributed ledger on a peer-to-peer network that is open, public, and anonymous. And blockchain is that underpinning technology that maintains the Bitcoin transaction ledger. 

How is Blockchain revolutionizing traditional business processes? How Bitcoin can help businesses grow?

One of the many ways in which Bitcoin can help a business grow is by enabling automation and efficiency in transactions. With the presence of Bitcoin, businesses have the ability to complete transactions quickly and seamlessly. Crypto allows businesses to use algorithms that allow financial transactions to occur in real-time. The barrier breakthrough with


now allows businesses to skip the complexity of traditional financial transactions on the internet as well as allow global access to various cash exchanges. Integrating cryptocurrencies like Bitcoin helps businesses to stand apart in the competition. With decentralization at hand, business leaders can complete transactions and funds without the obstacles of involving third parties.

What to use to successfully transform the business infrastructure?

5 Ways To Utilize Chatbots For A Better Customer Experience

An opted-in, highly engaged audience that enjoys real-time service from preferred brands will set the foundation for the ROI that businesses need to see

There’s no denying that chatbots drew in a lot of attention last year. The tremendous hype around their capabilities fueled intrigue about how they were going to change consumer behavior and produce impressive results. However, creating any chatbot experience – let alone a great one – is a much harder task than expected.

Everlane used its Messenger chatbot as an email alternative, sending messages such as order confirmations and shipping information. It fell victim to a rushed creation process and never had the time needed to generate real ROI because the holistic customer experience was not considered.

Download our Premium Resource – AI and machine learning for marketing

The purpose of this guide is to cut through the hype and noise around these powerful technologies and show what you can put in place today to boost your business results.

Access the AI and Machine Learning for marketing

Ultimately, many of today’s chatbots see substantial investments only to end up suffering the same fate. Chatbots shouldn’t be written off, though. With a strategic approach, they bring efficiency and ease to customer care solutions and give brands a new way to differentiate, strengthen loyalty, and increase revenue.

A challenging solution to build

It’s unrealistic to expect chatbots to solve all inquiries from the outset and even after processing huge quantities of data, research indicates that many reach a point of diminishing returns due to how they are built. While many chatbots languish in success rates, some of the best-equipped chatbots can accurately resolve 85% of inquiries, so it makes sense to maximize their effectiveness by pairing a chatbot with human personnel to resolve the remaining queries in collaboration.

Although many companies have struggled with internally developed chatbot solutions, leaders are pushing the boundaries beyond solving inbound inquiries and are now finding opportunities to proactively engage the customer and offer a more balanced blend of service, engagement and sales. This evolution is in its early stages and represents a significant leap ahead of the earliest chatbots.

5 ways to deliver exceptional customer experiences with chatbots

By equipping chatbots with the ability to collect and analyze data in real time, such as customer profiles and the context of inquiries, companies are creating a better and more engaging customer experience across service and engagement. In order to implement this new kind of chatbot to create measurable results for your business, follow these five steps:

1. Start by offering utility value

Provide customers with purchase-related information and alerts to build an opted-in audience on conversational channels, such as SMS and Messenger. This audience can then be used for engagement and upsell efforts, so the larger it is, the better. According to a 2023 report from Ubisend, 40 percent of customers want chatbots to send them special offers and deal alerts. In addition to sending purchase incentives, remind customers about forgotten items in their shopping carts, and offer them value-added services.

2. Put Data at the Center

AI-driven chatbots must serve multiple channels of interaction across a broad range of services and the limiting factor to this is data. Data enables a highly accurate contextual understanding of the customer and what he or she needs. Forrester research suggests that major companies such as Nike and Target are moving away from email as a channel for customer service needs in favor of real-time communications, including chatbots.

This change reflects the evolution in customer expectations and behavior, but to make it a reality, a rich customer data profile is required to support real-time automated assistance. When considering an automated assistant initiative, look at how data in the current ecosystem can be shared with the chatbot’s AI. Then, consider how all of the interactions and contextual understanding the AI has of a customer can be passed back into the existing data ecosystem. This roadmap will connect the dots regarding where data is and where it needs to go for chatbot success.

3. Understanding Context Is Key

Use industry-specialized AI to offer the customer more than a keyword-based, decision tree type of experience. Additionally, a chatbot initiative needs significant investment around planning the customer experience and offered services. In order to deliver on this vision without risks, utilize AI and related platforms that are proven and quick to deploy.

4. Move Beyond the Decision Tree

Conversation flowcharts are incredibly limiting and they aren’t effective at engaging customers. A flowchart approach bases a chatbot’s response on keywords in a customer’s query. For example, a bot might send a shipping confirmation if a user’s request has the word “delivery” in it. Essentially, the bot is limited by the capacity of the developers to anticipate use cases that the bot will serve.

The best chatbot experiences today are powered by AI that has a contextual understanding of the customer and is equipped with a wealth of data and capabilities to serve specific situations. Use platforms that not only provide the required conversational management and natural language understanding but also understand the context of the customer. It is equally important to verify that existing sources of data are made available to the chatbot in real time because a nightly refresh of data is no longer adequate.

5. Think Outside the Chatbot Box

Chatbots aren’t just notification tools — they’re a brand’s automated assistant. Naturally, they need to have a broad range of capabilities, from pre-purchase assistance to post-purchase customer care. Make sure to include other functions such as marketing, upselling and offering value-added services to the customer that create the differentiation your brand needs to thrive.

One report predicts that 70% of chatbots will be retail-based in just five years thanks to their ability to upsell, detect and prevent cart abandonment, as well as market loyalty and rewards programs that lead to satisfied customers who return for another purchase.

While initial efforts might concentrate on one specific channel or platform, it is important for brands to recognize the compounding value of offering their branded assistant across many channels, such as SMS, chat apps, voice platforms and on-site live chat. In every channel, the assistant should be able to provide contextual assistance, make the most of that channel’s potential and give the customer a seamless experience throughout the shopping lifecycle.

Chatbots have a long way to go before we see ubiquitous adoption, but researchers and developers are making it easier to launch and expand an automated assistant’s capabilities. Using sophisticated chatbots is a significant opportunity that offers the potential to revolutionize the customer experience and deliver differentiating services that secure loyalty.

To get there, leaders need to focus on the five steps above in order to give customers what they’re looking for. At the end of the day, an opted-in, highly engaged audience that enjoys real-time service from preferred brands will set the foundation for the ROI businesses need to see.

Better Know A Fix: Ddt

Welcome to another ongoing series on Our Modern Plagues: Better Know a Fix. It is the sister series to Better Know a Plague, which I introduced last week. The “fix” posts will explore the ways we use science and technology to thwart various modern plagues. Sometimes, I’ll line the posts up to match a specific plague I’ve already written about. Other times, that won’t be possible because there is no fix.

Fix is a funny word. To fix is to put something in order, or to make it more stable or permanent. Solving a problem fixes; it mends or repairs. But you can also be in a fix, which is akin to being in a quandary or a tight spot or hot water or any other such cliché.

Our first example is DDT, more formally dichloro-diphenyl-trichloroethane, which is widely credited with knocking down last week’s subject, the bed bug, after World War II.

DDT was the first modern synthetic pesticide. Paul Herman Müller, a Swiss chemist, discovered its insecticidal properties in 1939 after spending several years spritzing a glass box full of blue bottle flies with hundreds of chemicals. He was looking for a new insecticide with residual powers, which means that an insect that walked over or alit upon a treated surface would die. DDT worked so well that even after Müller washed his equipment, the chemical clung to the glass and continued to kill. Eventually, he had to dismantle the glass box, sanitize it, and air it out for a month to make it usable in new tests.

Müller’s main target was the Colorado potato beetle, an invasive species eating up crops on Swiss farms. His compound worked against this beetle. It also worked on houseflies and gnats. And since DDT’s discovery coincided with WWII, it wasn’t long before Allied forces were using it to control blood-sucking insect vectors, most famously malaria-carrying mosquitoes and typhus-carrying lice. Thanks in part to DDT, WWII was the first major engagement where fewer American troops died from disease than from weapons. Müller was subsequently awarded the Nobel Prize in Physiology or Medicine in 1948.

After the war, chemical companies offered DDT commercially, and soon we had DDT sprays, paints, wallpapers, dog powders, and more. DDT was on farmland and lawns, in orchards and homes.

To say we were overzealous with DDT is an understatment, and ultimately it caused long list of environmental and health problems, which you can read all about in Rachel Carson’s _Silent Spring _(although in retrospect, DDT was a relatively safe pesticide–its problems stemmed mostly from overuse). By 1972, the newly formed Environmental Protection Agency banned the use of DDT in the US, although we continued to manufacture and export it until 1982. The last American DDT plant, just outside of Los Angeles, is now a Superfund site.

Before its ban, the pesticide’s success was partly due to its novelty. Insects had never experienced such an assassin. Our previoulsy pesticides were mostly poisonous botanicals, or elements such as arsenic and mercury. None killed insects with such a pointed attack. DDT’s longevity was also an asset: it stayed on surfaces far longer than its predecessors, so was guaranteed to zap insects for a longer period of time.

The thing about insects is that they exist in droves and they are very, very prolific. Some are apt to have genetic mutations that let them dodge a pesticide—perhaps one that changes the shape of the ion channel, for example, preventing DDT’s deadly grasp. Genetic mutations flow quickly through generations of insects, and thus those naturally resistant will beget more resistant insects, which will do the same, ad infinitum.

A major example for DDT resistance involves a mutation dubbed kdr, or knock-down resistance, named for the fact that such mutant insects are difficult or impossible to knock down. Bed bugs started showing resistance within a few years of DDT’s widespread use, as did a long list of other insects including lice, mosquitoes, houseflies, fruit flies, and cockroaches.

Environmental impact aside, DDT’s other unintended legacy is that it functions much in the same way as a modern class of pesticides called pyrethroids. Nearly every over-the-counter insecticide in your cabinet contains a pyrethroid, as do professional-grade sprays, topical lice and scabies creams, and pesticide-impregnated clothes and bedding. The genetic mutations that made bed bugs and other insects resistant to DDT are the same that have made them resistant to pyrethroids, which has contributed to the massive bed bug resurgence we experience today.


Additional reading

Paul Herman Müller biography, Nobel Prize website

DDT – A Brief History and Status, EPA website

Nobel Lectures in Physiology or Medicine: 1942-1962_,_ World Scientific Publishing Co.

DDT and the American Century, David Kinkela

Widespread distribution of knockdown resistance mutations in the bed bug, Cimex lectularius (Hemiptera: Cimicidae), populations in the United States, Zhu et al, Archives of Insect Physiology and Biochemistry, 2010

12 Google Ads Extensions For Better Results

What are Google Ads ad Extensions? 12 Google Ads Extensions for Better Results Structured Snippet Extension Sitelink Extensions

A sitelink is an ad extension that’s widely used in all kinds of industries and business categories. It’s one of the oldest forms of ad extension, and it can help you connect with your potential customers by displaying additional pages of your website. For instance, if your ad is directed to a landing page, but your sitelinks take users to a different page that provides additional value or information, then that’s a good thing.

Location Extensions

A location extension is a type of app that lets people easily find their current location or a map of their current location. It can also include a call button or a phone number. This extension can be used for businesses that rely on in-person transactions, such as restaurants, barber shops, and beauty salons.

Affiliate Location Extensions

An affiliate location extension is a type of website that allows merchants who sell their goods through major retail chains to show their nearest stores. This helps search engines find you and your products.

Setting up an affiliate location extension is relatively easy. Just select the Google database of auto dealers or general retailers that you want to target. It will then show the closest proximity of these establishments to the searcher.

Product Extensions

Product extensions are helpful tools that allow you to enhance your Google Merchant account listing. They can also be used in campaigns that target keywords.

Call Extension App Extensions

Extensions for Android and iOS are great if you want to show off your app alongside your ad text. They should be especially important if you believe that the app helps drive site purchases, or if you see numerous transactions happening through it.

Seller Rating Extensions

Building trust and showcasing your business’ reputation are two of the most important factors that you can consider when it comes to building a strong online reputation. With seller ratings extensions, you can easily see how your business has been compared to other businesses. These ratings are then combined with the number of reviews to create a single rating.

Price Extension

Have you ever been a victim of an oil change price rip-off? I was able to avoid getting ripped off by a mechanic shop that tried to do it with me ahead of time, as I saw an extension for the price. Price extensions are not about preventing employees from deceiving their customers, but rather setting clear expectations. When users see an extension for the price, they are more likely to make an informed purchase.

Image Extensions Lead Form Extensions (New)

A lead form extension from Google Ads allows users to submit their contact details directly to the search results page. This eliminates the need for them to fill out an application form on your site’s landing page. If the user is a Google account holder, the relevant data can be pre-populated.

Automated Ad Extensions

You can’t manually add these extensions to your ad list. Google automatically chooses the appropriate ones for you based on its evaluation of how they can improve the performance of your ad.


Before you start using an extension, think about its purpose. It’s important to consider the quality of the work that it does and avoid extensions that are not helpful or do not contribute to your main goal. For instance, if you’re planning on using a lot of sitelinks, make sure that they’re not only good for you, but also for the development of your site.

Update the detailed information about Partial Auc Scores: A Better Metric For Binary Classification on the website. We hope the article's content will meet your needs, and we will regularly update the information to provide you with the fastest and most accurate information. Have a great day!