clickmodels 1.0.1

Probabilistic click models of user behavior on search engine result page

# [ClickModels](https://github.com/varepsilon/clickmodels)

ClickModels is a small set of Python scripts for the user click models initially developed at [Yandex](http://company.yandex.com). A *Click Model* is a probabilistic graphical model used to predict search engine click data from past observations.
This project is aimed to deal with click models used in Information Retrieval (see next section) and intended to be easy-to-read and easy-to-modify. If it's not, please let me know how to improve it :)

If you are using this code for your research work, consider citing one of our papers when appropriate
(see [References](https://github.com/varepsilon/clickmodels/#references) section below).

If you are looking for a general-purpose framework to work with probabilistic graphical models you might want to examine [Infer.NET](http://research.microsoft.com/en-us/um/cambridge/projects/infernet/). It should also work with IronPython.

# Quick Start

- `cp clickmodels/config_sample.py config.py`
- `vim config.py`
- `python bin/run_inference.py < data/click_log_sample.tsv 2>inference.log`

More details about the config and input data formats below.


# System-Wide Install
If you wish, you can install the ClickModels core (parameter inference and click simulation) to a system-wide location:

sudo python setup.py install

Uninstall:

sudo pip uninstall clickmodels


# New!
Now, thanks to [agrotov](https://github.com/agrotov), the models can also be run in a click generation mode and predict relevance (DBN only). Check out `ClickModel.get_model_relevances()` and `ClickModels.generate_clicks()` methods.

**N.B.**: Use this code with care as it is not fully tested yet.

***

# Models Implemented
- *Dynamic Bayesian Network* ( **DBN** ) model: Chapelle, O. and Zhang, Y. 2009. A dynamic bayesian network click model for web search ranking. WWW (2009).
- *User Browsing Model* ( **UBM** ): Dupret, G. and Piwowarski, B. 2008. A user browsing model to predict search engine click data from past observations. SIGIR (2008).
- *Exploration Bias User Browsing Model* ( **EB_UBM** ): Chen, D. et al. 2012. Beyond ten blue links: enabling user click modeling in federated web search. WSDM (2012).
- *Dependent Click Model* ( **DCM** ): Guo, F. et al. 2009. Efficient multiple-click models in web search. WSDM (2009).
- *Intent-Aware Models* ( **DBN-IA, UBM-IA, EB_UBM-IA, DCM-IA** ): [Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013).](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search)

***

# Format of the Input Data (Click Log)
A small example can be found under `data/click_log_sample.tsv`. This is a tab-separated file, where each line has 7 elements. For example, the line `1dd100500 QUERY1 50 0.259109 ["http://1", "http://2", "http://3","http://4","http://5","http://6","http://7","http://8","http://9","Http://10"] [false, false, false, false, true, true, false, false, false, false] [0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0]` has the following fields:

1. `1dd100500` — some identifier (currently not used)
2. `QUERY1` — text of the query. It can contain any UTF-8 characters except tab sign `\t`
3. `50` — integer identifier of the region (country, city) of the user who submitted the query. **If you don't want this, just put some constant (e.g., `0`) in this column**.
At Yandex user region is heavily used by ranking, so throughout the code the pair `(query, region)` is used to identify the query, i.e., if we have the same query string from two different region we consider them as two separate queries.
4. `0.259109` — float value, corresponding to the probability `P(I = V)` that user has a *vertical* intent V, i.e., he or she is interested more in the vertical documents than in organic web documents. **If you do not want any of this intent stuff just put `0` in this column**. We assume that the user has one of the two intents: *vertical* intent V with probability `P(I = V)` and regular *web* intent with probability `1 - P(I = V)`. For example, if we want to take into account user's interest in images and we somehow know the probability that the user is interested more in images than organic *web* results, we can make use of intent-aware click models. See [Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013)](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search) for more details.
5. **json** list of the URLs of the documents that make up SERP (search engine result page). Document's url is an identifier, so in principle you can use any (string) id you want. **NB**: this is not a python list, so yuo have to use double quotes and no comma after the last element.
6. **json** list with the *presentation types* (vertical types) of the documents (see [Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013).](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search)). **If you do not want to know this** just set it to the list of `false` of the same length as the previous list.
7. **json** list of clicks. Each element is the number of times corresponding URL was clicked

If you need more data to experiment with you can use any publicly available dataset and convert it to the format described above. For example, you can use a dataset provided by one of the Yandex challenges (you need to register to get access to the data):
- http://imat-relpred.yandex.ru/en/datasets
- http://switchdetect.yandex.ru/en/datasets

***

# Files
## README.md
This file.

## LICENSE, AUTHORS, CHANES.txt
Self explaining.

## setup.py
Python package installation file: `python setup.py --help`

## bin/
Directory with the scripts.

## clickmodels/
Directory with the core code. This is the directory that gets installed by `setup.py`.

## data/
`data/` directory contains an example of click log (see format description above) as well as two examples of result pages with fresh block included (see `makeGluedSERP.py` description above): `data/serp_sample.json` is used in an example above, while `data/serp_sample2.json` was used to create a picture in the paper [Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013).](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search)

## doc/
Automaticaly generated documentation. Also available [online](http://htmlpreview.github.io/?https://raw.github.com/varepsilon/clickmodels/master/doc/html/index.html).

## html/
Directory with the CSS/JS files for `bin/generate_serp.py`; it outputs there a number of HTML files for you to examine to get an idea what kind of SERPs we address with the intent-aware click models.

## bin/generate_serp.py
**{not used by the other, does not use other code}**
Create html of the SERP containing fresh block item. This is used just to illustrate the notion of *presentation types* used by Intent-Aware models. Run as: `bin/generate_serp.py < data/serp_sample.json`. Output is placed in `html/` directory.

**WARNING:** all previously generated html files in this directory will be removed


## bin/compare_click_models.py
**{not used by other scripts}**
Script used to compare different models and output significance of the difference. The pair of models to compare is specified in the code by modifying `TESTED_MODEL_PAIRS` variable. Model pair is a text string which is mapped to the pair of functions returning model objects (see `MODEL_CONSTRUCTORS` dict for the mapping). E.g. `UBMvsDBN` is used to compare UBM model (`UbmModel()`) to the default DBN model (`DbnModel((0.9, 0.9, 0.9, 0.9))`). **NB**: we may have a list of models needed to be compared to each other. For this purpose the same notion of *pair* is abused. For instance, `MODEL_CONSTRUCTORS['EB_UBM']` contains 3 algorithms to be compared to each other: **UBM, EB_UBM, EB_UBM-IA**.

- **Usage**: `bin/compare_click_models.py directory_with_click_logs 2>run.log`
- **Input**: `directory_with_click_logs` — directory containing files with click logs. Each file is in the format described above. These files are then sorted alphabetically and split into pairs where first file is used for training, the second one is used for testing. For example, if the directory contains files `f01`, `f02`, `f03`, `f04` then `f02` will be used to test models trained using `f01`, `f04` will be used to test models trained using `f03` and so on. Two models are evaluated on the test set and their performances (Average Perplexity or Log Likelihood) are compared using appropriate formula (see `perpGain` and `llGain` functions respectively). **NB**: Multiple train and test files are needed to calculate confidence interval for the gains (`bootstrap.py` is used for this purpose).
- **Output**: some progress output is printend to `sys.stderr` which might be useful for a long run. Finally the gains of one model over another is output in the following format:

UBM (0, 1) [-0.0115, 0.0659, 0.075, 0.0778, 0.0623, 0.0403, 0.0593, -0.040] (0.0095, 0.0662)
It first outputs the name of the "pair" `pairName` specified in the `TESTED_MODEL_PAIRS`, then pair of indeces `(i, j)` which mean that the model compared are `MODEL_CONSTRUCTORS[pairName][i]` and `MODEL_CONSTRUCTORS[pairName][j]` which is UBM and UBM-IA in our example. Next is the list of gains of model `j` over model `i` for each pair of the train/test logs (in our example we had 8 pairs of files under `directory_with_click_logs`). The next element is the confidence interval according to bootstrap test (with 95% confidence level and 1000 bootstrap samples). This line will be printed for all the "pairs" listed under `TESTED_MODEL_PAIRS` and for both Average Perplexity (PERPLEXITY) and Log Likelihood (LL). For perplexity measure also the gains for individual position (rank) are printed. Like this: `UBM POSITION PERPLEXITY GAINS: (0, 1) [[average_gain_for_pos1, confidence_interval_for_pos_1], …]`

## bin/run_inference.py
Run click model inference and evaluate the models:

- **Usage**: `bin/run_inference.py < data/click_log_sample.tsv 2>inference.log`
- **Input**: click log in the format described above (`sys.stdin`)
- **Output** (assuming that `TRAIN_FOR_METRIC = False`): `ModelName (LogLikelihood, Perplexity)`


## clickmodels/inference.py
This file contains implementation of all the click models, probabilistic inference and helper functions needed to work with them. This is the core of the codebase. More details about the classes/functions below.


## clickmodels/bootstrap.py
Copyright © [Ernesto P. Adorio](http://adorio-research.org/wordpress/?p=12295): code we use to perform a bootstrap test.


## clickmodels/config_sample.py
Copy this file to `config.py` and modify it; it will be used by `bin/run_inference.py` and `bin/compare_click_models.py`.
This is the file with default settings for different parameters of the inference, input and output data.

- `MAX_ITERATIONS` — maximum number of iterations in Expectation Maximization (EM) algorithm (applicable only for models using EM algorithm).
- `DEBUG` — perform some additional tests when running algorithm (makes it slower)
- `PRETTY_LOG` — make log output prettier. If `False` then more information is put into log.
- `USED_MODELS` — list of model names to be tested in `__main__` section of the script. Possible names are `['Baseline', 'SDBN', 'UBM', 'UBM-IA', 'EB_UBM', 'EB_UBM-IA', 'DCM', 'DCM-IA', 'DBN', 'DBN-IA']`. Please refer to the `__main__` section of `inference.py` to see how these names are expressed in terms of our class hierarchy (all those nasty `if 'XXX' in USED_MODELS`).
- `MIN_DOCS_PER_QUERY`, `MAX_DOCS_PER_QUERY` – number of documents per query. Set to 10 by default as most of search engines return list of 10 doucments.
- `SERP_SIZE` - size of the search engine result page (SERP). Used if we want to model clicks beyond the first result page. See the section named **Beyond the First Result Page** below for more details.
- `EXTENDED_LOG_FORMAT` - if set to `True` the urls, layout and clicks are dicts instead of lists (see **Format of the Click Log** section above). Example: `data/click_log_sample_extended_format.tsv`.
- `TRANSFORM_LOG` - transform the click log by inserting the fake documents for pagination button (currently works only with `EXTENDED_LOG_FORMAT = True`). See the section named **Beyond the First Result Page** below.
- `QUERY_INDEPENDENT_PAGER` - used to switch between `SDBN(P)` / `SDBN(P-Q)`. Only used with `TRANSFORM_LOG = True`.
- `TRAIN_FOR_METRIC` – if `True` the model will be trained such that its parameters can be used in a metric (like [Chuklin, A. et al. 2013. Click Model-Based Information Retrieval Metrics. SIGIR (2013).](http://ilps.science.uva.nl/biblio/click-model-based-information-retrieval-metrics)). See the section below for more details.
- `PRINT_EBU_STATS` — if `True` the parameters of the EBU metric will be printed first (*Yilmaz, E. et al. 2010. Expected browsing utility for web search evaluation. CIKM. (2010)*).
- `DEFAULT_REL` - default (prior) relevance (attractiveness, satisfaction) parameter values used in click models like DBN or EBU.

## clickmodels/input_reader.py
A class used for reading input data in a click log format described above.

***

# Class Hierarchy
Also see epydoc-generated [documentation](http://htmlpreview.github.io/?https://raw.github.com/varepsilon/clickmodels/master/doc/html/index.html).
## Click Models

The base class for all the click models is the class called `ClickModel`. In order to define a new click model you should create a clas inherited from it and re-define methods `train` and `_getClickProbs`.

- `train` function

Note, that `test` method is already implemented and uses `_getClickProbs` function. If you redefine `__init__` method, then be sure to invoke the `__init__` of the parent class to set the `ignoreIntents` and `ignoreLayout` parameters (they should be set to `True` unless you are using *Intent Aware* model)

`ClickModel` class by itself represents a baseline click model which sets probability 0.5 to any click event.

### DbnModel
This class is, in fact an implementation of a more general **DBN-IA** model ([Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013).](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search) ) that makes use of intent and presentation type of the documents when `ignoreIntent` and/or `ignoreLayout` is set to `False`. The `train` method is a probabilistic EM inference.

If all what you want is just original DBN model by Chapelle et al. you should creat it as `DbnModel((0.9, 0.9, 0.9, 0.9))` (`ignoreIntent` and `ignoreLayout` is `True` by default).

### SimplifiedDbnModel (DbnModel)
This is the same as `DbnModel((1.0, 1.0, 1.0, 1.0), ignoreIntents, ignoreLayout)`, but `train` method is just counting instead of EM algorithm. See *Chapelle, O. and Zhang, Y. 2009. A Dynamic Bayesian Network click model for web search ranking. WWW (2009).*, Section 5 (Algorithm 1).

### UbmModel
This is the most general case for all UBM-like intent-aware models. Changing `ignoreIntents`, `ignoreLayout` and `explorationBias` parameters you can get different models: **UBM, UBM-intent, UBM-layout, UBM-IA, EB_UBM, EB_UBM-intent, EB_UBM-layout, EB_UBM-IA** (for the names see [Chuklin, A. et al. 2013. Using Intent Information to Model User Behavior in Diversified Search. ECIR (2013).](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search)).

### EbUbmModel (UbmModel)
Just a shortcut for `UbmModel(ignoreIntents, ignoreLayout, explorationBias=True)` which correspond to the model called *Exploration Bias UBM* in *Chen, D. et al. 2012. Beyond ten blue links: enabling user click modeling in federated web search. WSDM (2012).*


### DcmModel
This model is, again, more general **DCM-IA** model which reduces to **DCM** when `ignoreIntents = True`, `ignoreLayout = True`. `train` method is a simple counting, no EM algorithm.

Please note, that `getGamma` method invokes `DbnModel.getGamma`, so be careful when changing that.


## InputReader
This class intented to read input (click log) in the format described above. To save memory, it maps queries and urls to ids. It means, that you need to use the same instance of the `InputReader` class even if you read multiple click log files. Otherwise you will end up with two different ids assigned to the same query.

***

# Performance Issues
If you experience performance issues consider using [PyPy](http://pypy.org/) instead of regular cPython. It may lead to 10x spead up.
You can also install and use [simplejson](http://pypi.python.org/pypi/simplejson/) module instead of `json`.

***

# TRAIN_FOR_METRIC
If you set `TRAIN_FOR_METRIC = True` the code will expect you to provide document relevances instead of URLs. We make an assumption, that document attractiveness and/or satisfaction probability only depends on its human-assigned relevance grade. A model will then be trained to assign the same attractiveness / satisfaction probabilities to all the documents with the same relevance.

## Input Format
The format in this case is similar to the one descirbed above with only difference that URLs should be replaced by the relevance grade of the corresponding document to the corresponding query. The query field will be ignored in that case. The relevance grade should take one of the following values:

- `IRRELEVANT` — lowest relevance scorn, document is not relevant to the query
- `RELEVANT` — combines marginally relevant and just relevant documents
- `USEFUL` — document is more than just relevant, it is really useful
- `VITAL` — highest relevance score, the document is essential

Please note, that if you also have `PRINT_EBU_STATS` set to `True`, then the parameters of the EBU / rrDBN metric will be printed out first (these ones can be computed directly without need to train a model).

## Output
For each model corresponding parametres will be printed out:

- `UBM` — attractiveness probabilities `alpha` and position discount parameters `gamma`
- `DCM` — attractiveness probabilities `alpha` (named as `urlRelevances` in the code) and position discount parameters `gamma`


For more conceptual details about converting click models into evaluation metrics please refer to the paper [Chuklin, A. et al. 2013. Click Model-Based Information Retrieval Metrics. SIGIR (2013).](http://ilps.science.uva.nl/biblio/click-model-based-information-retrieval-metrics)

***

# Beyond the First Result Page
If you want to model the clicks beyond the first result page you may want to model pagination button separately. We implemented the models described in the paper [A. Chuklin, P. Serdyukov, and M. de Rijke. Modeling Clicks Beyond the First Result Page. In CIKM. ACM, 2013.](http://ilps.science.uva.nl/sites/ilps.science.uva.nl/files/cikm1325-chuklin.pdf). Namely, by setting the following config options you will get:

- `TRANSFORM_LOG = True`, `QUERY_INDEPENDENT_PAGER = False`: `SDBN(P)` model
- `TRANSFORM_LOG = True`, `QUERY_INDEPENDENT_PAGER = True`: `SDBN(P-Q)` model
- `TRANSFORM_LOG = False`: reqular `SDBN` model

Please, refer to the paper for more details.
***

# References
- *A. Chuklin, P. Serdyukov, and M. de Rijke. Using Intent Information to Model User Behavior in Diversified Search. In ECIR, 2013.* [[pdf]](http://ilps.science.uva.nl/biblio/using-intent-information-model-user-behavior-diversified-search)
- *A. Chuklin, P. Serdyukov, and M. de Rijke. Click model-based information retrieval metrics. In SIGIR. ACM, 2013.* [[pdf]](http://ilps.science.uva.nl/biblio/click-model-based-information-retrieval-metrics)
- *A. Chuklin, P. Serdyukov, and M. de Rijke. Modeling Clicks Beyond the First Result Page. In CIKM. ACM, 2013.* [[pdf]](http://ilps.science.uva.nl/sites/ilps.science.uva.nl/files/cikm1325-chuklin.pdf)

# Copyright and License
Copyright © [Yandex](http://company.yandex.com) 2012-2013, [varepsilon](https://varepsilon.github.io) 2012-∞

Published under the BSD license.