In day-to-day life, a highly demanding task for IT companies is to find the
right candidates who fit the companies' culture. This research aims to
comprehend, analyze and automatically produce convincing outcomes to find a
candidate who perfectly fits right in the company. Data is examined and
collected for each employee who works in the IT domain focusing on their
performance measure. This is done based on various different categories which
bring versatility and a wide view of focus. To this data, learner analysis is
done using machine learning algorithms to obtain learner similarity and
developer similarity in order to recruit people with identical working
patterns. It's been proven that the efficiency and capability of a particular
worker go higher when working with a person of a similar personality. Therefore
this will serve as a useful tool for recruiters who aim to recruit people with
high productivity. This is to say that the model designed will render the best
outcome possible with high accuracy and an immaculate recommendation score.
Data is examined and collected for each employee who works in the IT domain focusing on their performance measure.
データは、IT領域で働く各従業員のパフォーマンス指標に注目して収集される。
0.71
This is done based on various different categories which bring versatility and a wide view of focus.
これは多目的性と焦点の広い視野をもたらすさまざまなカテゴリに基づいて行われる。
0.80
To this data, learner analysis is done using machine learning algorithms to obtain learner similarity and developer similarity in order to recruit people with identical working patterns.
But there is a greater percentage of companies who do not follow up on their employees regarding their skills.
しかし、従業員のスキルについてフォローしていない企業は、より多く存在する。
0.64
Moreover, it has become difficult to assess the quality of the candidate and whether they would fit right in with the company culture.
さらに、候補者の質や企業文化に適合するかどうかを評価することは困難になっている。
0.65
A lot of variables and factors come into consideration regarding the quality of a candidate the company wants to hire and these vary from different companies concerning different job positions.
Predominantly, the quality of a candidate depends upon different factors such as the candidate’s perks, competency, CCI (Continuous Capability Improvement), participatory culture, and so on.
Also, there is a vast set of competition between the candidates solely over the chances of going to a higher rank in their company.
また、社内で上位に進む可能性についてのみ、候補者同士の競争が激化している。
0.63
In these types of scenarios, there comes a comparison metric between the preceding candidates who are trying to move for a higher rank.
この種のシナリオでは、上位に進めようとしている前任者の比較基準が提示される。
0.53
To address these issues, this research provides a new metric system for the companies to help in identifying the right set of candidates for employability within their company.
2 Literature Survey [5] Employability is a major factor in the sustainable competitiveness within the company as well for the personal growth of the candidate.
2 文献調査 5]採用性は,企業内における持続的競争力と,候補者の個人的成長の主要な要因である。
0.76
As a result, a question arises in the mind of everyone - Is the candidate chosen the best option from the pool of choice?
Companies have a hard time answering the said question due to a lack of monitoring and bias in the hiring process itself.
企業は、雇用プロセス自体の監視とバイアスが欠如しているため、この質問に答えるのに苦労しています。
0.68
[9] Implicit bias towards characteristics like race, gender, and sexual orientation has always resulted in persistent inequality and reduces the utility of the employer.
Let us consider two users A and B. If A and B are interested in the same product, then their preferences for other products may align more compared to any random user.
A と B が同じ製品に興味があるなら、A と B の製品に対する好みは、どのランダムなユーザーよりもよく一致しているかもしれない。
0.69
In the modern data-driven world, recommendation systems play a huge role in supporting the process of decision-making.
The hashtag-based clustering method with a combination of several rounds of the k-means algorithm and the DBSCAN algorithm is used to form clusters of data that belong to like-minded people.
Chaotic data of millions of people can be filtered out easily and working partners for each worker can be assigned by the hirer thus increasing the efficiency of both the workers.
[1] Academic credentials include but are not restricted to diplomas, degrees, certificates, and certifications, which act as a way to attest completion of training or education undertaken by the student.
Broadly speaking, these credentials may also attest to the successful completion of any test or exam.
広く言えば、これらの資格証は、テストや試験の完了を成功させる可能性がある。
0.62
Ultimately, they serve as a model of independent validation of the said individual’s possession of the knowledge, skill, and ability needed to carry out a particular task or activity.
The most competent skills that are meant to be present among the candidates that are proclaimed by larger companies are based on Kaggle, Github score, etc., Information regarding one’s Kaggle or Github helps the company to identify the level of skill-set of the candidate.
Insights such as the quality of contributions made by a developer on GitHub could provide an initial understanding of a potential candidate’s coding capabilities.
Considering the rising popularity of social recruiting and the kind of potential GitHub platform offers, GitHub can be leveraged for software developer recruitment.
They generate a complete profile from their profile analysis of those candidates who submitted their resumes.
彼らは履歴書を提出した候補者のプロフィール分析から完全なプロフィールを生成する。
0.74
Avrio has its own Fitscore for each candidate to compare them based on various factors and the generated profile will be run through every job offering database to see if there are any matches.
Also, they have their very own virtual assistant called Rio which is responsible for the candidates’ personal profile as it is programmed to record and present the analysis of the candidate by asking different sets of questions that will reflect their skills, capability, and interest set for the available job.
The required parameters will be made by the clients so that the algorithm consistently searches those third-party websites for any matching for the candidates.
The selected pair consists of non-negative real numbers.
選択された対は非負の実数からなる。
0.59
The Canberra distance is used for comparing ranked lists and used in computer systems to check for any malicious activities or any violation of policy.
In simple words, it is the sum of the absolute difference between all dimensions of two points It is extensively used in a vast array of fields from regression analysis to frequency distribution.
5 Proposed Method To overcome this problem, the model tries to find the similarity between candidates.
5 提案手法 この問題を解決するために、モデルは候補間の類似性を見つけようとする。
0.75
In this paper, a new standardization metric is introduced for measuring a candidate’s potential skills on the technical scale for the recruitment process from a company’s point of view.
5.1 Turtle-Score A score based on the learning analytics score, GitHub score, Puzzle meter, Job Shadowing meter, and error archive score is the metric that is used to measure one’s potential skills in the technical scale for the recruitment process.
Github score is the sum total of all these outcomes and it mainly focuses on the number of commits that a particular person had contributed to all the repositories over a span of time.
This shows the consistency and the working nature of a person within a few seconds thereby saving time.
これは、数秒以内に人の一貫性と作業性を示し、時間を節約します。
0.73
5.3 Learning Analytics Turtle Score
5.3 学習分析のタートルスコア
0.64
LA score is derived from Learning Analytics which is based on the collection and analysis of different websites and articles that the candidate had come across while working.
Learning Analytics is an extension that helps in collecting and storing articles, research papers, websites, etc., which seems to be helpful for a person to solve an error or to complete the coding process.
This score can help in determining the interests and skill set of a candidate in a particular area of knowledge by analyzing all the data collected by the individual using the Learning Analytics extension.
5.4 Kaggle Score Kaggle is a well-known website for Machine Learning candidates which offers datasets and Jupyter Notebooks to publish and run ML models.
Errors occupy an important position in the career of a person as it is a part of the learning process and it helps change the perspective of viewing a particular problem.
During this period, if the candidate is capable of understanding what exactly is going on and if they are able to work on the said technology without any guidance, then they get a high score.
This is used to see where a person stands on their learning and observation scale.
これは、人が学習と観察の尺度でどこに立つかを見るために使用される。
0.71
5.7 Puzzle Meter
5.7 パズルメーター
0.64
Puzzle meters are an exclusive tool to identify candidates with an urge to solve problems given to them no matter how long it takes or how hard the task is.
A candidate can be assigned with tasks from different areas of expertise thereby the Manager can assess them by considering the time taken to solve the particular problem and the interest of the candidate in that particular task which made them work off the clock.
Thus, deserved candidates can be chosen for the jobs accordingly.
したがって、適格な候補者をその職に選べる。
0.68
5
5
0.42
英語(論文から抽出)
日本語訳
スコア
Therefore by combining all these scores, a person would be able to determine the total skill-set of a candidate and also find similar candidates with the same talents.
Flask framework in this model takes in a user’s details as input and provides with similar candidate list as output along with its similarity percentile.
The difference among each candidate’s data is calculated and an overall similar percentile score is found.
各候補のデータの違いが計算され、全体として類似したパーセンタイルスコアが見つかる。
0.72
7.1.5 ML Integration
7.1.5 ML統合
0.56
This includes the prediction of similar candidates using the related similar score accordingly using the machine learning tools and the accurate results are produced.
これには、機械学習ツールを用いて類似したスコアを用いて類似候補を予測し、正確な結果を生成する。
0.73
7.2 Data Architecture
7.2 データアーキテクチャ
0.71
Figure 2: Architecture Chart
図2:アーキテクチャチャート
0.77
7.2.1 Training Flow
7.2.1 トレーニングフロー
0.57
The inputs for the training model will be loaded.
トレーニングモデルの入力がロードされます。
0.69
The input data set that has been loaded for the training model is the company dataset of all candidates.
トレーニングモデルのためにロードされた入力データセットは、すべての候補の企業データセットである。
0.85
Now, the data is prepared for training where it will undergo some process to fill null values and remove outliers in the provided data set.
These results are calculated based on the turtle score.
これらの結果はカメのスコアに基づいて計算される。
0.63
Therefore the comparison is done based on this and similarity is then calculated accordingly.
したがって、これに基づいて比較を行い、それに応じて類似度を算出する。
0.72
It depends on how closely they are related, that is, the smaller the difference between the scores of any two candidates, the closer is the relation between them.
9 Future Works Over this course of Turtle score research, the comparison is done using Distance Similarity measures such as Euclidean, Manhattan, Minkowski, and many more.
There are over 9 different Distance similarity measures used in this research and further research can be done with many more similarity measures.
この研究には9以上の異なる距離類似度尺度があり、さらに多くの類似度尺度で研究を行うことができる。
0.80
Further research can also be done by bringing up many more entities.
さらなる研究は、さらに多くのエンティティを持ち上げることで行うことができる。
0.73
This might lead to a much closer similarity.
これは、より近い類似性をもたらす可能性がある。
0.67
Also, the addition of different unique entities makes the turtle score more compatible and customizable as organizations from different sectors can use it as well.
Data can also be derived from various candidates belonging to various age groups.
データは様々な年齢グループに属する様々な候補者から得られる。
0.80
This will provide the dataset with diversity which in turn will improve the application range of the model.
これが提供します 多様性のあるデータセットは、モデルの適用範囲を改善します。
0.71
10 Conclusion The primary ideology behind this applied research is that the process of recruitment is made to be efficient in order to hire candidates like whom exactly the company wishes to have.
This in turn encourages students to stay on their feet and improve.
これにより、生徒は足元に留まり、改善することが奨励される。
0.66
Finding similar candidates not only makes it possible to recruit people based on the requirements but also provides more data about a person as similar data is stored in the dataset.